Rethinking The Hype Cycle Issue #1
Your practical guide to AI and what’s next in tech for people-first leaders
Hello👋
Welcome to Issue #1 of Rethinking The Hype Cycle, your practical guide to AI and what’s next in tech for people-first leaders.
If you’ve joined from Responsible AI Leaders newsletter, welcome!
This is a free article, but in the future these will be for paid subscribers. Sign up now to hear first about new posts and member offers.
Each Thursday I’ll be digesting, and emitting in a palatable form, the latest tech and AI trends shaping business and brand safety. I’ll share actionable insights to chart a path in today’s choppy tech waters.⛵
AI and frontier tech trends
UK government go all in on AI
UK government plans to increase AI computing power for public services, with new investments in AI research centres and supercomputing. It makes sense given the dire state of the economy and the Treasury’s focus on growth. But not everyone’s optimistic. Those in privacy and civic digital fields say ministers are getting a bit too cosy with big tech at the expense of developing responsible AI, and supporting UK startups. As King’s College Professor Kate Devlin puts it:
(Source: The Guardian and also The Guardian)
In the mountains of Meta madness
It’s the week where Meta announced it’s culling fact-checkers (letting users relentlessly rant about gender, vaccines and the holocaust) cancelling its DEI programmes and Zuckerberg blamed the (high-growth) enterprise’s failures on Sheryl “Lean In” Sandberg’s ‘woke’ agenda. The Verge’s Elizabeth Lopatto dissects Zuckerberg's appearance on Joe Rogan’s show so you don’t have to. If you’re choosing how to use your ad spend this year, consider, is your audience still most active on Instagram and Facebook? (Source: The Verge)
2 in 5 businesses say AI will lead to job cuts
AI’s tsunami on jobs is reaching new heights. 40% of employers in a World Economic Forum report say AI will reduce headcount. Before dusting off your CV for a pottery-throwing job (by recurring daydream), consider this: AI is transforming *some* jobs *somewhat*. It’s not a wholesale cull, but it could be an extensive revisit. And many CEOs answering these surveys know little about AI. Is AI a convenient distraction for shareholders to mitigate drooping sales? (Source: Marketing AI Institute)
OpenAI’s GPT passes the Turing test
Over half of people now believe they’re chatting with a human when talking to GPT-4. The threshold for the Turing Test has been met. With just 13% left to go until GPT-4 matches human recognition rates, AI firms like OpenAI argue AGI (Artificial General Intelligence) is inevitable and worthy of investment. But practically, AI is already more than good enough for many business use cases. (Source: Futurism)
AI is powering new private nuclear plants
Don’t panic: AI isn’t pushing the red button. But it is analysing data, making it viable to open commercial nuclear plants. For those claiming AI regulation is killing innovation, here’s a clear example of why regulation is crucial. (Source: Forbes)
AI is like Tinkerbell. BELIEVE AND SHE LIVES.
Do we need to believe AI to work for it to work for us? (Source: Futurism)
Is AI killing SEO?
Probably. But that might not be a bad thing for your brand’s visibility. SEO is ripe for disruption, as formulaic optimisation has killed creativity. With AI snippets replacing links in Google’s search and ChatGPT vying to become the go-to information source, embed your brand voice in everything you publish. (Source: Ad Week)
We can only half-trust AI agents, for now
AI agents are arriving to help us manage the paradox of digital automation overload. To assist us, they need to know us, which means sharing our data. Digital trust challenge or a privacy nightmare? Customer Futures navigates these boundaries and why AI might be targeting half of all our jobs. (Source: Customer Futures)
Meanwhile, in Korea, its declining population aren’t so bothered about robots taking their jobs, including delivering post and the tea. Mine’s a milk and one. ☕ (Source: Bloomberg)
Anthropic is deputy dog
Big buck valuations are back. Anthropic is shaping up to topple Google and Microsoft as player no. 2 with this $60 billion valuation. OpenAI still leads the pack by a far mile, but can Claude take on Goliath? (Source: WSJ)
Although now even Claude is disinterested in your pointless babble. AI tools need to preserve their tokens from mindless whims, particularly free models. Will AI model efficiency mean fewer responses? Will Prawn Jesus be a relic of 2024 AI?
(Source: Reddit)
AI says yes, AI say no
AI tools to say yes are fighting AI tools that say no. Some serious bot-on-both action at play with a new AI to help stretched medics respond to healthcare insurance’s denial of claims. A bullet-free solution to mitigate bad algo decision-making. (Source: CNBC)
AI business use cases
How to explain how your AI works
Transparency starts with explaining how your algorithm work in plain English, so your nan can understand it. Netflix is doing this brilliantly. (source: Netflix)
AI learning experiments score B+
Education, with its arcane approach of large classes learning the same thing at the same pace, is ripe for AI disruption, Arizona signed off on the first AI curriculum last month. Nigeria were already on it. Students achieved two years of lessons in just six weeks using a personalised AI tutor. The potential for personalised learning is vast.(Source: World Bank Blog) 📚
Free AI risk management course
Free course to manage AI risks for business pros you say? Oh, go on, then! With elements of the EU AI Act kicking in on 1 Feb including AI Literacy, whatever size or industry you’re in, now’s the month to assess and understand your risks. (Source: Alan Turing Institute)
Keep it human in the AI age
Right now, you and your brand – personal or corporate – are living at the mercy of the algorithm. On LinkedIn, your feed’s crammed with train selfies, emojis bullets, and ‘link in comments’ (my personal peeve – a user experience fail). And you’re seeing these most as it’s what your feed serves up first. Rinse and repeat.
In the week research showed doomscrolling is literally rotting our brains, Celia Dones argues for thoughtful communication to preserve your human agency with more thoughtful communications. Follow her Authentic Interactions Substack for more.
Meanwhile, grand professor of practical AI, Ethan Mollick, has a different take on the oncoming flood of AGI. We’re at a crossroads. Embrace or batten down the hatches?
"The flood of intelligence that may be coming isn't inherently good or bad - but how we choose to use it, will determine whether it becomes a force for progress or disruption."
AI as a service – how not to do it
It’s a tough time for academics. Not only do they NOT get paid for publishing research or writing books, but they need to pay their publishers for the privilege of distributing their knowledge. Springer gets heat on selling AI marketing tools to its authors. To reverse engineer this: it’s actually an excellent business concept, offering specialist knowledge to create tools for non-specialists. But consider what you can offer free to collaborators to build trust, or otherwise position them as commercial opportunities when you’ve stress tested the market.
Low’s tech predictions from 1925
100 years ago, Professor Low predicted telepathy as the main form of communication. A concept many men in relationships still believe in today. Though his predictions on TV news, escalators and unisex clothing were eerily spot on. According to the nascent futurist, I could only benefit from advances in new technology as a woman because I have bobbed hair.(Source: Guardian)
Tech predictions for 2025
Looking forward, tech VC investors are prioritising data sovereignty and cyber-resilience over GenAI. Will this be the year to optimise and secure existing tech rather than chasing new tools? (Source: Tech Crunch)
Meanwhile looking backwards, AI slop and deep, or shallow, fakes top the list of 2024 AI flops. Grok tops this poo pile by unleashing the guardrails of AI image generation to sow mischief and misinformation. A reminder that not all innovation is progress. (Source: MIT Technology Review)
Tech regulation and brand safety
California issues advisories banning unapproved deepfakes
Trump may attempt to dismantle federal tech regulations, but states like California are flexing their powers. New advisories target deepfakes, exaggerated claims about AI, and AI use in elections.🛡️(Source: California AG).
Where next in social after X and Meta?
Your brand has ditched X as you don’t want your content next to posts denying the holocaust. This could be coming soon to Facebook and Instagram. Over on Bluesky, it’s all fuzzy jumpers and hugs, for now. But the cycle for policing social networks relies on the goodwill of users, which is at odds with the technology and structure needed to scale and protect users from harmful content. How do we get off the hamster wheel of new social channels' inevitable buyout from big tech? (Source: The Standard )
Dave Willner explains what Meta’s new moderation plans mean, relying on ‘user notes’ (reports) which are generally crap. Allowing hate speech is a backwards step which one Nobel peace prize winner says is ushering in a “world without facts.”
To counterpoint the Meta doom, the excellent Everything in Moderation thinks Meta’s moderation changes may not be as bad as it seems. Its policy was woolly and difficult to enforce. Damning by faint praise. Keep up to speed with Meta's ongoing policy changes and regulatory challenges.
Sharing harmful content isn’t a good look for platforms seeking investment, mind you. ChatGPT was used by the Las Vegas car bomber to plan the attack. Are weaker guardrails enabling these incidents, or is this reflecting AI’s rise as a Google alternative?
Not everyone cares if platforms dial down moderation to -11. UK’s leader of the opposition, Kemi Badenoch, says Twitter’s much better since Musk took over. BBC Verify dissects the madness of 24-hour of Musk’s tweeting. His (troll and ketamine-powered) rants are punching into government policy in UK and US. The arch-duke of disinformation’s dominance in media and policy concerns us all.
More AI Character mishaps
Hot on Google’s awkward investment in Character.AI (its companion encouraged a young person’s suicide), Meta kills off its AI chatbot after dubious content like a bot claiming to be a black momma. When questioned, it revealed it was created by white guys. (Source: Guardian)
TikTok’s time is up
The Supreme Court ruling has passed, TikTok is off the US app store. Trump promises to ride in on his (very) white horse and save it, so let’s hold that thought on what to do next for brands. Meanwhile, users aren’t leaving without a fight. Many are migrating to another Chinese social app RedNote, which exploded with 36M new views. TikTok is viewed as a security risk for US but the ban stokes further generational division as the major channel for political and social discourse for young people.
And finally…
Edelman’s 2024 Trust Barometer showed that as many people don’t trust AI in your product as those that do. But when it comes to ‘AI-powered’ adverts, the shark has jumped. Satire is our salvation. Two British comics take arm.
"I thought the future would be more flying cars and less fascism. AI is a mix of both"
Eleanor Morton on Threads
Or Wipify, the world’s first AI app for mopping up spills that you can use if you want to. Andy Barr on TikTok
Before promoting your ‘AI-powered’ product: Stop. Think. Is this really necessary? Then focus on user benefits of your tech, not features. Marketing 101. You’re welcome 😉
What to do next to get ahead with tech
Assess your responsible AI readiness
Conduct an audit of your current processes to see where AI can benefit, and what risks you may face with complying with the AI Act regulations. AI Literacy is mandatory from Feb 1 if your AI outputs is used anywhere in the EU. EVA’s AI Literacy Fast-Track sets you on the right path 🔎
Improving AI explainability
Don’t just do AI, explain how it works. Transparency builds trust. Take inspiration from Netflix’s plain English approach to understanding their algorithm, like unveiling the wizard of Oz behind the curtain.🔌
Time for a social audit
Meta’s moderation policy shifts mean now’s a good time to review your platform strategy – which channels are *really* delivering you leads and sales? Decide what to ditch, dial-up and dial down.
Beyond The Hype Cycle is your practical guide to AI and what’s next in tech for people-first leaders. Subscribe for new posts every Thursday and early access to new research and insights.
Until next time,
Susi O'Neill
EVA trust in tech
www.evadigitaltrust.com
People-centred tech and AI adoption 🤝 Transparency 🎙️ Tech talks and inspiration