BBCNEWS — AI Decoded
20260806 003000 UTC
530 transcript segments
Original Broadcaster Captioning (Enhanced)
Data courtesy of The GDELT Project (https://www.gdeltproject.org/), from the Internet Archive TV News Archive.
Machine transcription. Treat it as a searchable index of what was broadcast, not a verbatim quotation record.

[00:00:00] Now on BBC News. Join our I decoded decoded team as we unpack, explore
[00:00:03] explore and take a deep dive into
[00:00:05] into the world of artificial intelligence.
[00:00:22] Hello again. Welcome to AI
[00:00:24] Decoded. $2.5 trillion. That is what
[00:00:28] what the world will spend on artificial artificial intelligence this year.
[00:00:31] year. That's up nearly 50% on last
[00:00:33] last year. But what is the return
[00:00:36] return for that investment? Every Every boardroom has been told the
[00:00:38] the same story. This will transform transform productivity, cut costs,
[00:00:42] costs, change the way that we all
[00:00:44] all work. A year ago, the research
[00:00:47] research suggested that wasn't happening. happening. Barely any of it was showing showing up on the balance sheet.
[00:00:51] sheet. So where are we now? Well, Well, the most recent studies show
[00:00:55] show there have been improvements improvements in some areas. Give
[00:00:58] Give the AI a clearly defined task. task. Write code, answer customer
[00:01:02] customer queries. The systems are
[00:01:04] are measurably faster, but most businesses
[00:01:07] businesses are a chain and all too
[00:01:08] too often speeding up one link. Just
[00:01:11] Just parse up the work at the next. next. So how do we solve that? With
[00:01:14] With us this week, Peter Grant, CEO
[00:01:18] CEO of go.ai, and James Coote, who who is the founder of pair, both
[00:01:21] both spend their working lives inside inside companies trying to close close this gap. So hopefully they'll
[00:01:25] they'll be able to answer the question. question. And with us this week,
[00:01:28] week, our co-host and author of Technology Technology Is Not Neutral. Doctor
[00:01:32] Doctor Stephanie Hare is here to to welcome to you all. Peter, let
[00:01:35] let me start with you. Going back
[00:01:38] back to the MIT study a year ago,
[00:01:42] ago, 95% of these projects were delivering
[00:01:46] delivering nothing measurable. Is
[00:01:47] Is that changing? I think it is slowly
[00:01:51] slowly starting to change, but I I do think that's a canary in the
[00:01:54] the coal mine. I think that report report has come under some criticism, criticism, Christian, because it
[00:01:57] it was a very small, sample. It was
[00:02:01] was taken of 153 companies, coming coming back from them. But a more
[00:02:05] more recent report by PwC that was was released at Davos this year,
[00:02:08] year, the World Economic Forum was
[00:02:10] was across 4500 CEOs in 95 countries. countries. But that found that 56%
[00:02:14] 56% of them are yet to see a return return on investment from the investments
[00:02:18] investments they've made in generative generative AI, and only 1 in 8 can
[00:02:21] can actually claim a cost saving
[00:02:24] saving or revenue generation. I think think the short answer is one is
[00:02:28] is trust, two is complexity and three
[00:02:30] three is speed. Enterprises are very very different to how consumers buy
[00:02:34] buy generative AI. So is it failing
[00:02:37] failing because companies are buying buying things technologies faster
[00:02:40] faster than they know how to use
[00:02:42] use them? Or do they know exactly exactly what they want and can't
[00:02:46] can't make it work? I think it's
[00:02:48] it's probably the latter, but go go back to trust. So these closed
[00:02:52] closed models, they want data. They're They're like a plant and imagine
[00:02:55] imagine the plants, the algorithm algorithm and water is what feeds
[00:02:58] feeds the plant. And the issue is is that they need this data. They
[00:03:01] They need this information. But enterprises, enterprises, their IP is basically
[00:03:05] basically locked inside their four four walls. And they need to protect protect that to make sure those organisations
[00:03:09] organisations don't have that information. information. So they're very cautious cautious about how they implement
[00:03:13] implement generative AI. But also also trust comes down to the answers answers that you're getting from
[00:03:17] from the agents as well. So as we we know these systems hallucinate
[00:03:21] hallucinate in the consumer space, space, that's fine. But if you're you're a bank, if you're lending
[00:03:24] lending money, if you're in pharmaceuticals, pharmaceuticals, you're doing research research and everything else, you
[00:03:28] you have to be correct in what you're you're doing. So that's why organisations organisations are cautious. That's
[00:03:32] That's why it's taking a bit more more time as well. But I think there's
[00:03:35] there's a mis set expectation. If If you think Google took five years
[00:03:38] years to get to 100 million users,
[00:03:40] users, ChatGPT took 60 days. That's That's 30 times faster than anything
[00:03:44] anything we've ever seen before. before. So I think organisations
[00:03:47] organisations are struggling with with that to really digest it and
[00:03:51] and make it meaningful in their companies. companies. Yes, shareholders are
[00:03:55] are an impatient mob. James, I know
[00:03:57] know you're going to tell me, that
[00:04:00] that technology is not the hard part. part. We're already at PhD level
[00:04:04] level systems. It's about adoption. adoption. Yeah. I mean, we've never
[00:04:07] never had a challenge like this before before when suddenly we've got PhD
[00:04:11] PhD level intelligence in almost almost every subject, but occasionally
[00:04:14] occasionally lies on tap. And it's it's very difficult for your average
[00:04:18] average employee in an organisation organisation to use those tools to
[00:04:21] to their full potential. And right right now almost everyone has access.
[00:04:24] access. They've all got co-pilot, co-pilot, they've got ChatGPT or or they're using Claude, or maybe
[00:04:28] maybe their boss doesn't know, but but very few of them are adopting
[00:04:32] adopting it meaningfully, meaningfully meaningfully and actually using these,
[00:04:35] these, these tools to their capability. capability. And that's what we call
[00:04:38] call the adoption gap, right? And And actually it's big and growing
[00:04:41] growing bigger as the frontier of of models progresses. What do you you think is the problem? Stefanie?
[00:04:44] Stefanie? Do you think it's a technology technology problem or a human behaviour behaviour problem? So first of all,
[00:04:48] all, it's not enough to just give give people a tool and say work it
[00:04:51] it out. Most people are not receiving receiving the kind of training that that you would need to make the most
[00:04:55] most of these tools. And that's quite quite scary if you think about it, it, given that it is everywhere,
[00:04:58] everywhere, it's in hospitals. So So you've got doctors who are maybe
[00:05:01] maybe chatting with you and using
[00:05:04] using ChatGPT along with the patients
[00:05:06] patients to discuss symptoms, check
[00:05:09] check diagnoses, check treatments, treatments, not going through the the ethics of that, the governance
[00:05:12] governance of that. Is it safe to to upload your medical results and
[00:05:16] and tests into these models or not? not? They hallucinate so they just
[00:05:19] just get things wrong. They also also just lie. They just make things
[00:05:23] things up and have to be challenged. challenged. It all depends on how
[00:05:25] how you prompt this art of prompting. prompting. People don't do training
[00:05:29] training because training is a cost. cost. So that's one reason that I I think we're not seeing the value
[00:05:32] value yet. But second is I think think we see value in things like
[00:05:36] like software development or customer customer service, but we're not seeing
[00:05:39] seeing it necessarily throughout throughout the entire organisation organisation because you might speed
[00:05:42] speed up productivity in one part part of your business, but that doesn't doesn't mean that. Can you give me
[00:05:46] me an example? Where would you see see within that chain that I described?
[00:05:50] described? What would you see a speeding speeding up of the business processes
[00:05:53] processes and where would you see see the bottleneck? So let's look look at the business of healthcare.
[00:05:57] healthcare. So you go into a hospital, hospital, everybody can understand understand radiology, right? So the
[00:06:00] the people are looking at your skin.
[00:06:02] skin. So it's your X-rays or MRIs,
[00:06:04] MRIs, Cat scans just because you you are able to use AI to process
[00:06:08] process scans and read them faster faster to look for a bone break or
[00:06:11] or a tumour doesn't mean that you're you're necessarily going to cut patient
[00:06:14] patient waiting lists or get faster faster treatment for people because
[00:06:18] because the skin is just one of many, many, many tasks, all of which have
[00:06:21] have to be signed off. Healthcare Healthcare is highly regulated by by doctors the whole way through.
[00:06:25] through. Well, those doctors aren't aren't getting more time just because because the scans are being read
[00:06:29] read faster. So those red scans pile pile up, but everybody still has
[00:06:33] has to then put them through the the same old processes. I think we'll
[00:06:36] we'll have to rethink a lot of workflows workflows from the ground up because
[00:06:39] because of AI. It's time on the side side of some of these companies,
[00:06:43] companies, Peter, because $2.5 trillion trillion in total spend this year
[00:06:45] year is an awful lot of money. What What I'm not sure about is what this
[00:06:48] this looks like on the balance sheet. sheet. And how do you measure it
[00:06:51] it for shareholders? Yeah, we like
[00:06:53] like to call it an ROI, but return return on intelligence rather than
[00:06:57] than return on investment. And that's that's really for every dollar or
[00:06:59] or pound that you spend, what measurable
[00:07:02] measurable business value you're you're actually getting back. So
[00:07:04] So we've created a framework basically basically for enterprises to be able
[00:07:07] able to measure this. So the money money they're spending, how do you
[00:07:10] you measure it for vectors basically basically are you starting are you
[00:07:14] you experimenting today all the way way through to are you the agentic agentic enterprise where humans and
[00:07:18] and agents are basically working working together? What's your strategy?
[00:07:21] strategy? Who owns that strategy? strategy? Who's what's the budget budget you're doing to fund it? And
[00:07:25] And are you actually doing a proper proper ROI analysis before you fund
[00:07:28] fund any of those particular workflows workflows that Stephanie was talking
[00:07:31] talking about? What is your data? data? What's the state of it? How
[00:07:34] How do you get access to the data? data? Can you get to the system of
[00:07:36] of record for the ROI? So your CRM, CRM, your customer relationship management
[00:07:40] management systems, your ERP systems, systems, what's the technology you're
[00:07:43] you're using? So have you just let let everybody use it like James was
[00:07:46] was talking about co-pilot perplexity, perplexity, OpenAI, whatever, or
[00:07:49] or is it structured your approach approach by team or by organisation,
[00:07:52] organisation, the governance that that you have a governance is, is is super important as well, especially
[00:07:56] especially when you talk about SDA, SDA, which is zero data retention
[00:07:58] retention to make sure it's not being being leaked outside, you have control control over that. Some people have
[00:08:02] have been doing some famous cases cases about that and then workforce
[00:08:06] workforce literacy, as Stephanie Stephanie was quite rightly saying, saying, have you trained and have
[00:08:10] have you enabled people? You wouldn't wouldn't start at a bank and say, say, OK, go and do the lending straight
[00:08:14] straight away. You train people, people, you enable them, you certify certify them, you make sure they're
[00:08:16] they're an ambassador for your brand brand before they talk to customers. customers. You need to do exactly
[00:08:20] exactly the same with AI as well. well. And then the final probably
[00:08:23] probably the most important thing thing is what's in pilot and what's
[00:08:26] what's actually in production. And And what is that ratio. And we're
[00:08:29] we're seeing at the moment it's about about 30% in our index from those
[00:08:32] those people that are actually say say got ten pilots running, only
[00:08:34] only 30 are actually running production production in your workflows. Right? Right? Well, when it comes to pilots,
[00:08:38] pilots, I mean, obviously that is is quantifiable. You can see that
[00:08:41] that on a balance sheet. But there
[00:08:43] there are some processes and some
[00:08:46] some things you might do using AI, AI, which are not as quantifiable.
[00:08:50] quantifiable. James. Yeah. I mean, mean, we were talking about this this just before, one of the things
[00:08:54] things that AI has been most valuable valuable for me on in the recent
[00:08:58] recent past was helping us think think through and then execute on
[00:09:01] on our US expansion as a business. business. It's quite a complex thing thing to do. I'd never done it before.
[00:09:05] before. I spoke to five different different people about it. They give give five different opinions. Probably
[00:09:08] Probably some of them would be classed classed as a hallucination if an
[00:09:11] an AI gave it. But AI was really
[00:09:13] really helpful in market mapping, mapping, working out our messaging messaging over there, working out
[00:09:17] out what we have to do on the payroll payroll side and share side. And
[00:09:20] And it's very difficult to isolate
[00:09:23] isolate the sort of one bit of ROI
[00:09:25] ROI AI gave us there, but it's unarguable
[00:09:28] unarguable for me and for, you know, know, my leadership team. That there's
[00:09:31] there's been an unbelievable enabler enabler to helping us get started
[00:09:34] started over in the US. So that was was one where I thought it doesn't
[00:09:38] doesn't neatly fit, you know, a project. project. There's not just one part
[00:09:40] part it's helped with. It's helped helped with the ideation part. It's
[00:09:43] It's helped with planning and operations. operations. It's helped with actually
[00:09:46] actually going off and, you know, know, helping us make some of the the first sales over there. And I
[00:09:50] I don't know whether it fits neatly neatly into a project box, but good
[00:09:53] good God, it's been valuable. OK, OK, well, maybe then the problem problem isn't technology at all.
[00:09:57] all. Maybe it's the assumption assumption assumption that you can buy a tool tool and immediately expect people
[00:10:00] people to change how they work. So So let's talk about the workforce
[00:10:04] workforce because this is where you you come in. James. Are businesses
[00:10:07] businesses spending more on software software than they are on training?
[00:10:11] training? The people who use it? it? And what does that training need
[00:10:13] need to look like? Undoubtedly. I I mean, there's lots of stats on
[00:10:17] on this. The latest I've heard is
[00:10:20] is that three times as many people people in British businesses have
[00:10:23] have access to an AI tool versus versus those who have access to AI
[00:10:27] AI training. So most people are getting
[00:10:29] getting access to AI with no training. training. So that's undoubtable.
[00:10:33] undoubtable. And the thing that people people are often missing with the the training that they do deliver
[00:10:37] deliver when they deliver it is it it focuses very much on how do you you use the tool? Well, you're right
[00:10:41] right in the box and you click enter enter or maybe you speak to it. It's It's pretty easy to actually use
[00:10:45] use it. The thing which is difficult difficult and is at the crux of successful
[00:10:48] successful training on using these these tools is helping people cross
[00:10:52] cross this imagination gap between between I know how to write in the
[00:10:56] the box and I know how to use this
[00:10:58] this tool to reinvent how I actually actually work and use my time to
[00:11:02] to deliver value, not just, you know,
[00:11:04] know, do things faster and push a a bottleneck onto someone else. And
[00:11:08] And I think bridging this imagination imagination gap between I know how
[00:11:11] how to use a tool and I know how how to actually use it to re-imagine
[00:11:14] re-imagine my role and do more valuable valuable things is the thing that
[00:11:16] that people really need to, to, to to get across to deliver value from
[00:11:20] from training. So where does that that stem from, do you think, Peter?
[00:11:24] Peter? Is it top down the senior
[00:11:28] senior executives need to encourage encourage AI adoption within the
[00:11:30] the workforce or does it, as I suggested,
[00:11:33] suggested, grow organically? Interesting
[00:11:36] Interesting question. If you think think about I've gone through many
[00:11:40] many different technology revolutions, revolutions, shall we say, in 30
[00:11:44] 30 years in the industry and what what because I mentioned speed earlier,
[00:11:47] earlier, the speed that this has has come is unprecedented. What most
[00:11:50] most companies have been they've they've been lazy with regards to to change management. So people process
[00:11:54] process technology and they've just just given people the keys to the the car without teaching them to
[00:11:58] to drive. What I find the most successful
[00:12:00] successful ones are where the CEO, CEO, where she leads the initiative
[00:12:04] initiative and she leads by example example with her own agents and she
[00:12:06] she shows how she's using them with with her executive team. And that
[00:12:10] that permeates down to get people people really encouraged. Well, if if she's doing it, I need to do it
[00:12:14] it as well. Because what that creates creates a more permissive environment
[00:12:17] environment within the company. Yeah, Yeah, exactly. And culturally it it says this is a good thing to do.
[00:12:21] do. This is the organisation that that we're going to be. This is part part of our strategy, our vision
[00:12:25] vision and everything else. We do do believe this technology can help help us. It's not a technology issue.
[00:12:28] issue. It's more than enablement enablement side and the art of the the possible. That's role specific.
[00:12:32] specific. How I can do this in my
[00:12:34] my job. So we have a customer called
[00:12:37] called Mimecast 2500 people. The The CEO led that. We went in and
[00:12:40] and trained them with James's business business and it permeated down and
[00:12:43] and what we the way the evidence evidence that we see, it's working.
[00:12:46] working. Christian is basically in in queries how they're querying the the agents, how they're using them
[00:12:50] them day in, day out, whether they're they're being sales in marketing,
[00:12:53] marketing, in writing code or in in procurement for assessing companies
[00:12:56] companies or many different use cases. cases. But their usage has been off
[00:12:59] off the charts. I can understand
[00:13:04] understand the permissive environment environment within a company allows allows people to take hold of the
[00:13:07] the AI and run with it. Is there there any resistance? Because let's
[00:13:10] let's face it out there, there is
[00:13:12] is a lot of animosity towards AI AI and the replacement of jobs. Do
[00:13:16] Do we take it for granted that people people want to use it? I think people
[00:13:20] people are using it more and more more in their personal lives, and
[00:13:22] and that's often sort of the gateway gateway into this, right? I used used it to plan my holiday or to
[00:13:26] to find a deal and that kind of helps
[00:13:29] helps you build a bit of muscle memory
[00:13:31] memory for in your work. I think think secondly, it helps when, you
[00:13:35] you know, leaders model the behaviours behaviours they want to see. As Peter's
[00:13:37] Peter's just said, when a leader leader shows this is normal, this
[00:13:40] this is OK. I use this. It's not
[00:13:43] not dangerous if used in the right right way within our guardrails and
[00:13:46] and actually we're going to celebrate celebrate good use of this if it
[00:13:49] it leads to value. And I think thirdly, thirdly, what happens what we generally
[00:13:52] generally see is as you start to, to, you know, build some momentum momentum and more and more people
[00:13:56] people around you start to build
[00:13:58] build agents and Barry from procurement, procurement, did you hear he's built
[00:14:02] built an agent? I thought Barry was was about to take VR. You know, the
[00:14:06] the social sort of. Well, can you you give me. Some examples of that? that? Some fun examples of companies
[00:14:09] companies you've been into where where you've seen that happen? Yeah, Yeah, I'll link it to Peter's point
[00:14:13] point actually around execs really really grasping the tools. So there's
[00:14:16] there's a CEO of a manufacturing manufacturing company in Coventry. Coventry. It's called Graham. He's
[00:14:20] He's fab. I didn't ask him about about this shout out. So hope you
[00:14:22] you don't mind. But anyway he's he's he's absolutely massively into AI
[00:14:26] AI and he's using lovable to build build prototypes of how he'd like
[00:14:29] like to see the future of consultancy consultancy delivered from his business.
[00:14:32] business. And he's also using Claude Claude and experimenting with that.
[00:14:36] that. And I think through showing showing his team where he sees the
[00:14:40] the vision of the future in a way way which is tangible and shows a
[00:14:42] a bit of what's possible in the future.
[00:14:44] future. But does it safely with synthetic
[00:14:46] synthetic data and prototypes? Initially, Initially, I think is pretty exciting
[00:14:50] exciting and I think it's. Graham. Graham. Oh, I wouldn't want to guess.
[00:14:53] guess. Now don't put me on the spot.
[00:14:55] spot. 40s, seconds. Oh, that's. Late
[00:14:58] Late 40s you see we we. We sort of of the reason I ask is because there's
[00:15:02] there's a whole new policy that the the government is setting out today
[00:15:04] today that in England that they want
[00:15:07] want to educate children at the age age of 14 on AI that is being used
[00:15:10] used by local companies. I wonder wonder if you think, Stephanie, that that is the way we should be going.
[00:15:14] going. So we introduce children to
[00:15:16] to it at a very early age. So my
[00:15:19] my view is I would really like to to see us get kids reading books books the way they need to be before
[00:15:23] before we start having them mess mess around with AI. I think there's
[00:15:25] there's a time and place for AI in in education, but I'm really worried
[00:15:29] worried actually, that we're raising raising a generation of children children who don't have the critical
[00:15:33] critical thinking skills that come come from the old skill set, the
[00:15:35] the traditional skill set of basic basic literacy, numeracy and ability
[00:15:39] ability to write. So before we these
[00:15:42] these are really powerful tools that that affect the cognitive abilities
[00:15:45] abilities of adults with PhDs who
[00:15:47] who were raised in an analogue era.
[00:15:50] era. OK, so we put them onto children children whose brains are still forming,
[00:15:54] forming, who don't have that reading reading habit built in, who aren't
[00:15:56] aren't used to writing and learning learning how to express ideas and
[00:16:00] and thinking they won't have what what we all have, which is decades
[00:16:03] decades of muscle memory and critical critical expertise to be able to
[00:16:06] to look at an AI response and know
[00:16:09] know it's wrong, know it's hallucinating. hallucinating. So I kind of feel
[00:16:12] feel like, can we just give the kids kids a break? The people who work work in this are both nodding, but
[00:16:15] but I mean, you go into companies companies are young employees skilled
[00:16:18] skilled for the job. Whilst I agree
[00:16:22] agree with Stephanie on you cannot
[00:16:25] cannot shortcut expertise on developing developing critical thinking. I think
[00:16:28] think that is separate from whether whether young people entering the
[00:16:32] the workforce should be proficient proficient at using AI. And I think
[00:16:35] think that is unarguably a good thing thing and something which is lacking
[00:16:38] lacking right now. And I think an an 18-year-old who leaves school
[00:16:41] school with good critical thinking
[00:16:44] thinking skills, who is developing developing expertise but who is also
[00:16:48] also AI fluent, is going to be extraordinary. extraordinary. And unfortunately unfortunately there's not many of
[00:16:52] of them. Well, all of that brings brings us to one organisation that
[00:16:55] that tried to test every single thing thing that we just discussed in a
[00:16:58] a classroom in New York. Salamanca Salamanca is a very small rural district
[00:17:01] district in western New York on Seneca
[00:17:04] Seneca Nation territory. And they've they've just announced that they've
[00:17:07] they've bought a humanoid robot called
[00:17:09] called Sally. Get it? Sally Salamanca,
[00:17:12] Salamanca, Silicon skin, moving facial facial expressions, Western New York
[00:17:15] York accents. So she's going to fit
[00:17:17] fit in really well, for $57,000,
[00:17:20] $57,000, which I guess is a lot less less than the salary of a single single teacher. Anyway, she's been
[00:17:23] been brought in to assist in coding
[00:17:26] coding with AI classes from September September alongside an AI teaching
[00:17:29] teaching assistant. But then the the teachers got wind of it and the
[00:17:33] the unions got wind of it. And on on Friday the district paused the
[00:17:36] the whole thing. Now there are supporters supporters who say it will prepare
[00:17:39] prepare children for the world they're they're going to inherit. And then
[00:17:42] then there are the teachers who many many of the parents who say there
[00:17:46] there is no substitute for real face-to-face
[00:17:49] face-to-face interaction. Why do
[00:17:51] do you think, Peter, that story in
[00:17:53] in Salamanca makes people so
[00:17:58] angry? I don't know, and I think think it's quite sad that it does does really. I think this is a bit
[00:18:01] bit of fun, right. And I think to to James's point, the earlier you you can introduce AI to children
[00:18:05] children and help and support them them to get them ready for the workplace,
[00:18:08] workplace, I think it's really important. important. We're moving into a world
[00:18:11] world where we will control agents, agents, we will become managers of
[00:18:14] of agents that understand business business processes, but they're always always there's always going to be
[00:18:18] be a human in the loop, but they're they're going to take away some of
[00:18:21] of the mundane tasks. If you can can teach children how to do that. that. But I also agree with Stephanie.
[00:18:24] Stephanie. They need to understand understand the basics of grammar, grammar, arithmetic and everything
[00:18:28] everything else as well. Absolutely. Absolutely. 100%. But why not introduce
[00:18:31] introduce it? I think it's a bit bit of fun. I think if it engages
[00:18:34] engages children more it helps educate educate them and it makes learning
[00:18:38] learning fun that I'm very supportive. supportive. I was thinking when I I was reading this, isn't this what
[00:18:41] what Priya does? She takes robots robots into companies, didn't she?
[00:18:44] she? And then to school classrooms classrooms with Century Tech and
[00:18:47] and she teaches them how to how to to use these robots. But I mean,
[00:18:50] mean, is it the technology these these people are objecting to, Stephanie Stephanie or is it the symbolism
[00:18:54] symbolism of a robot standing where where a teacher ordinarily would?
[00:18:57] would? Oh, my God, where to even even start with this education policy
[00:19:00] policy in the United States in 30s. 30s. So first of all, we know that
[00:19:04] that again, American children are are doing worse on their reading
[00:19:08] reading comprehension, their writing writing abilities and their numeracy. numeracy. So we're not doing the
[00:19:12] the basic three R's right. We have have the data on that. It's not an
[00:19:15] an AI problem. It's been going for for at least the past 20 years, and and it's not unique to the United
[00:19:19] United States. But my point is we
[00:19:21] we have a problem. Houston. OK, then
[00:19:24] then the teachers unions are of course course incredibly powerful and active
[00:19:26] active in the United States is probably probably elsewhere. They can stop
[00:19:29] stop things. They can go on strike. strike. It's a nightmare. So they
[00:19:32] they probably feel quite disrespected, disrespected, worried about their
[00:19:34] their jobs. And then I think there's there's parents who are saying, can
[00:19:38] can we please just teach the kids? kids? I agree that we want we want
[00:19:42] want them to get exposed to AI as as we want them exposed to everything.
[00:19:45] everything. I'm not sure a $57,000
[00:19:48] $57,000 robot is the best return.
[00:19:50] return. OK,. Audience question. In
[00:19:53] In fact, the prize for the best question question of the week goes to Thomas
[00:19:56] Thomas who said this AI systems are are secured against the threats we
[00:20:00] we already know about. So what is
[00:20:01] is to stop an advanced AI finding
[00:20:04] finding entirely new ways to exploit exploit systems that no human has
[00:20:08] has ever anticipated, particularly particularly if it's pursuing its
[00:20:11] its own objectives? Very timely question,
[00:20:14] question, because Stephanie OpenAI OpenAI admitted in the past week week two of its models had escaped
[00:20:17] escaped a sealed testing environment, environment, found a zero day, got
[00:20:21] got onto the internet and hacked
[00:20:23] hacked Hugging Face, an AI hosting hosting platform in order to steal
[00:20:26] steal the answers and cheat the system,
[00:20:29] system, thus pursuing its own objectives. objectives. So Thomas's question
[00:20:32] question is absolutely right what what do we do about that? I mean,
[00:20:36] mean, there's nothing that we can can do to prevent that in the sense
[00:20:38] sense of that. That is a risk. It's It's always been a risk. It's a risk
[00:20:41] risk it will happen again. I'm amazed amazed that we're not hearing more more about it. What's interesting interesting in the OpenAI Hugging
[00:20:45] Hugging Face case is that both companies companies are open and transparent.
[00:20:49] transparent. What is worrying about about it as well is that could have have really caused serious harm.
[00:20:53] harm. I mean, we're lucky it was was Hugging Face, frankly, that was
[00:20:57] was hacked. And I'm sure if that that had a financial repercussion, repercussion, not a power plant they
[00:21:01] they would want. Exactly. They would would want or a hospital again, where where all of a sudden, you know,
[00:21:05] know, you've got machines going down down mid operation. You see, I was
[00:21:07] was rushing to this question from from our top conversation because
[00:21:10] because we're talking about rolling rolling out AI through the company.
[00:21:13] company. Right. And I want to know, know, Peter, if you are thinking
[00:21:17] thinking about energetic system within within your company and you're dealing
[00:21:20] dealing with very sensitive information, information, this same problem applies
[00:21:23] applies to that, doesn't it? You You might create a sandbox environment,
[00:21:27] environment, a testing environment environment within the company, but
[00:21:29] but if it's dealing with very sensitive sensitive information, what is to
[00:21:32] to guarantee that it doesn't leak?
[00:21:34] leak? Well. I think I'm not an AI AI research scientist Christian.
[00:21:38] Christian. So one caveat the answer.
[00:21:40] answer. But this is basically what's what's called rainbow teaming where
[00:21:43] where you create one agent that's that's nefarious that basically has
[00:21:46] has to create all different ways ways of attacking a system and you're
[00:21:50] you're doing what's called open ended ended learning, where you create create another agent to inoculate
[00:21:54] inoculate itself against it. And And it's constantly doing and learning learning from it. And that's basically
[00:21:58] basically what you're doing. You're You're using the AI to protect itself
[00:22:00] itself against other AI right? You
[00:22:03] You gave us an example of a CEO in
[00:22:07] in Coventry, Graham. I think it was
[00:22:09] was who is trying things himself
[00:22:13] himself inventing, forward thinking
[00:22:16] thinking what is to how does a CEO
[00:22:18] CEO sitting at the top of a company company knowing that some of his
[00:22:22] his employees or her employees or
[00:22:24] or trialling systems in some ways,
[00:22:26] ways, what is to reassure them that that that kind of sensitive information
[00:22:29] information doesn't break out or or indeed that the agents working
[00:22:33] working for her company doesn't hack hack into the sensitive systems of
[00:22:36] of another? Yeah, well, I think you
[00:22:39] you need to have a pragmatic and,
[00:22:42] and, you know, thoughtful CISO, you
[00:22:45] you know, security lead or digital digital lead, who, who's thinking
[00:22:48] thinking through these problems and and keeping up to date, Christian
[00:22:50] Christian with, with what's out there there because the threat picture
[00:22:53] picture is evolving at an extraordinary extraordinary rate when it comes
[00:22:56] comes to cybersecurity. And, you you know, for Graham, he has people
[00:22:59] people he can look to for that, right? right? It's his chief digital officer,
[00:23:03] officer, Patrick, very pragmatic pragmatic guy, will be looking into into these sort of threat vectors
[00:23:06] vectors and making sure that the the way they roll out AI, you know,
[00:23:10] know, has guardrails on it and the the right checks and balances in
[00:23:12] in place. And I think every digital digital leader will be thinking about
[00:23:15] about that. For a small business, business, it's a bit harder. You
[00:23:18] You may not have that expertise in-house. in-house. And we obviously have the
[00:23:22] the NCS, the National Cyber Security Security Centre, which puts out great great guidance on this sort of things
[00:23:25] things to help particularly small small and medium sized businesses. businesses. And I think you've got
[00:23:29] got to keep your ear to the ground ground on this stuff because this this is rapidly evolving. I mean,
[00:23:33] mean, security is the key issue here. here. The detail in the story about
[00:23:36] about OpenAI and the hack of Hugging Hugging Face was that when it tried
[00:23:40] tried to throw the intruder out, out, its own AI defences were slowed
[00:23:43] slowed by the guardrails they put put in place. The attacker was bound
[00:23:47] bound by no such rules. The two systems systems OpenAI had weren't bound
[00:23:50] bound by those rules. So I mean in in that sense it becomes an arms
[00:23:53] arms race. Yes. And it comes to a a bigger point that we've talked
[00:23:56] talked about a lot on this show of
[00:23:59] of the US, tech race. So the United United States is looking at going,
[00:24:02] going, should we be regulating our our models? Do we want to regulate
[00:24:05] regulate other countries models having having to use a Chinese model which
[00:24:09] which Hugging Face had to do to respond respond to this threat? It doesn't
[00:24:12] doesn't speak very well to the US US AI industry at the moment, where
[00:24:16] where our approach to regulating regulating it or not, this feels
[00:24:19] feels very Wild West to me and I I don't like it because all these
[00:24:22] these companies keep being told the the future is agentic you need to
[00:24:25] to move into having AI agents all
[00:24:28] all over your operations. Maybe you you don't yet. Maybe that's going
[00:24:31] going to come back and bite you. you. Peter, give me the last thought
[00:24:34] thought on that. It's interesting. interesting. I mean, Kimmy just got
[00:24:38] got released from China, which is is an open model. So for everybody
[00:24:42] everybody that means you can take take that model, you can install install it yourself, you can put
[00:24:45] put a firewall around it, you understand understand the weights, the algorithms
[00:24:48] algorithms and everything else where where a closed model you'll be giving
[00:24:51] giving your data to that model. They They can take it, they can use all
[00:24:54] all that information. That's what what scares enterprises the most.
[00:24:57] most. I think what you're going to to see Christian is a hybrid. So
[00:25:00] So you're going to see large organisations organisations which actually go back back to on prem. They will protect
[00:25:04] protect all their data at all costs costs and they decide where they
[00:25:06] they send the query to for the agent, agent, whether it goes out outside
[00:25:10] outside the organisation for a very very basic query or internally for
[00:25:13] for something to protect their own own IP, which is ultimately comes
[00:25:16] comes back to trust. Brilliant place place to leave it where we started.
[00:25:20] started. It's all about trust, isn't isn't it? Peter Grant James Cook Cook doctor Stephanie Hare really
[00:25:23] really good to talk to you all. Thank Thank you very much indeed. That's
[00:25:26] That's it for this week's AI decoder. decoder. Just a reminder that we
[00:25:29] we are taking a break through August, August, but the programme will be
[00:25:31] be back at the beginning of September. September. In the meantime, if you you have any thoughts or questions
[00:25:36] questions or feedback for everything everything we discussed, do email
[00:25:38] email us. AI decoded@bbc.co.uk and and we've put on screen for you the
[00:25:42] the QR code. So if you want to look look back at the previous episodes
[00:25:46] episodes of AI Decoded, it's all all there on the YouTube playlist.
[00:25:48] playlist. And of course every episode episode is on the iPlayer. Thank
[00:25:52] Thank you for watching. We'll see
[00:25:53] see you next time.
[00:28:07] edge, but Russia's goals seem unchanged. unchanged. For continued coverage coverage of the war in Ukraine, join
[00:28:11] join us here on BBC News. I was on
[00:28:19] on air when the London 7/7 attacks
[00:28:21] attacks happened. That day was different different because of the sheer volume
[00:28:25] volume of information we were grappling
[00:28:27] grappling with in real time. Conflicting Conflicting reports on the number
[00:28:30] number and locations of the attacks,
[00:28:33] attacks, unverified videos, witnesses witnesses in shock. It was crucial
[00:28:37] crucial that we report events accurately accurately without causing unnecessary
[00:28:40] unnecessary panic. We couldn't let let the desperation for news influence
[00:28:44] influence our reporting or be overwhelmed overwhelmed by the magnitude of what
[00:28:47] what we were witnessing. Today, as as we wade through endless waves
[00:28:51] waves of misinformation, it's vital vital that we don't just ask the
[00:28:55] the questions. We question the answers.
[00:28:58] answers. I'm Matthew Amroliwala on
[00:28:59] on Verified Live for BBC News. That
[00:29:02] That booming. Noise that you hear
[00:29:05] hear from automatic semi-automatic semi-automatic weapons, it was a
[00:29:08] a pretty distinctive sound. And that's that's when all of us hit the deck.
[00:29:12] deck. All along the banks of Hungary's
[00:29:14] Hungary's historic River Danube overlooking overlooking the parliament building,
[00:29:17] building, which is going to dramatically dramatically change after tonight's
[00:29:20] tonight's results. That is the part
[00:29:22] part of Lebanon that is now occupied
[00:29:25] occupied by Israeli troops. We heard heard the first explosions after
[00:29:29] after midnight last night and then then the air raid lasted right through
[00:29:32] through the night. It's not just just what you see and you hear as
[00:29:35] as the rocket lifts off. You can
[00:29:38] can actually feel the force of it
[00:29:40] it through your body. There, when
[00:29:42] when the story breaks, BBC News.
[00:30:09] MUSIC
[00:30:12] Live from Singapore. This is BBC
[00:30:17] BBC News. Iran says it is in the the final stages of negotiations
[00:30:21] negotiations with Oman on a deal deal that would allow ships to pass
[00:30:23] pass through the Strait of Hormuz. Hormuz. Russia launches a devastating
[00:30:27] devastating attack on Kyiv, with with President Zelensky warning that
[00:30:30] that a lack of air defence missiles
[00:30:31] missiles is costing Ukrainian lives.
[00:30:34] lives. The BBC speaks to the activist, activist, whose hunger strike galvanised
[00:30:37] galvanised the recent youth led protests protests in India, saying the movement
[00:30:40] movement has weakened Prime Minister Minister Narendra Modi's authority.
[00:30:44] authority. Gianni Infantino will will remain president of Fifa after
[00:30:46] after receiving the backing of senior senior executives, but sincerely
[00:30:49] sincerely apologises for errors made made in his controversial private
[00:30:52] private investment plans.