🌟 Editor's Note
Welcome to the bi-weekly newsletter which focuses on the AI topics that leaders need to know about. In this AI age, it’s not the knowledge of AI tools that sets you apart, but how well they can be integrated in the context of your business.

If you find this newsletter insightful, please share it further !

Featuring

• Three Things That Matter

• In Case You Missed It

• Tools, Podcasts, Products or Toys we’re Currently Playing With

How Is AI Being Used by Consumers ?

• Menlo Ventures (a long standing Silicon Valley venture capital firm) have released a report sharing how consumers are using AI. 61% of Americans in the survey have used AI, with 1 in 5 people relying on it every day. As the authors point out, if these numbers were scaled up across the world, it’s little wonder the topic is so hot at the moment.

• Alongside this are estimates for consumer spend on AI tools - currently slated at $12bn. Side note - this is not to be confused with enterprise AI spend. But as 97% of consumers are using free versions, there is a lot of upside to be captured.

• In amongst the hype in the doc that we should anticipate as the VC world amps up the noise around AI to justify valuations, there are a few elements organisations should look to. For instance, is this adoption of AI by people individually also being captured by organisations - are they bringing this adoption into the workplace? And what is stopping the remaining 39% of people adopting - can this be addressed inside organisations to help companies capture more of the value that AI could generate for them?

• Generational adoption is also interesting - whilst GenZ are the biggest adopters (probably unsurprising to most), Millennials are power users, with the biggest adoption of daily usage. 45% of baby boomers are adopting AI - a figure that some other generations may not have felt to be the case

Why This Matters

Overcoming the barriers for adoption and helping organisations through the change - it’s hopefully reassuring that the #1 reason for avoiding change is ‘preferring people over AI’. This is good for humanity ! Our ways of collaborating as groups of people at work will not change quickly.

We’ve seen academic papers over the past 18 months that have shown that whilst individuals are using AI more, most organisations are not yet seeing the benefit of this change. For instance, a recent standout qualitative study demonstrated that it’s the organisational challenges and not people’s adoption that gets in the way. (Study by Thomas Übellacker can be found here https://arxiv.org/pdf/2502.15870).

This study points out that adoption of a new technology in the workplace evolves through cycles of individual understanding, social learning and organisational adoption.

• Individual understanding can be helped by exposure.

• Social learning is best addressed by teams focusing on how they can collectively adopt new technology into working practices, taking into account all the voices in the team. Those who resist as they prefer people can be helped to see that this technology is a tool and not a replacement colleague.

• The organisational barriers such as governance and providing access and training for AI tools can only be addressed by leadership: relying on teams to do this themselves is not a sustainable strategy.

• And the people who don’t see a need for AI may be the ones that you need to look most closely at inspiring - if this is a lack of information about the transformational nature of this technology perhaps you have a chance. If it is willful ignorance or a refusal to change, leaders will need to ask themselves whether these individuals hold back progress. As the Melo Ventures survey points out “people adopt tools that help them do what they already need to do, but in a better, faster, cheaper way”. If we help people in our organisations to replace routine tasks to create small wins, the value compounds and more people buy in.

What to watch out for next : predictions for consumer AI from the report that are worth being aware of

• AI models will become ‘social’ - very much in line with Ubellacker’s research: humans aren’t likely to change our ways of getting things done in groups quickly (such as approaches to hierarchy, the fact that we need to collaborate etc)

• Voice AI will take off (already we are seeing large numbers interact via Siri and Alexa

• Physical AI will enter the home (Superhuman.ai have a weekly robotics special which can keep you updated in this area)

• Revenue models go beyond subscriptions (ad models, affiliate revenue etc

Full Menlo Ventures report is here: https://menlovc.com/perspective/2025-the-state-of-consumer-ai/#ea137d0d-d6bc-4183-8a18-a5103d388b20

Answer me or I’ll kick the puppy ! Should you be nice to your AI ?

Some people are polite to their AI out of a cultural disposition. Others think that please and thankyous will get better results. Tipping was a commonly shared tactic for performance and Sergey Brin, one of the founders of the search engine Google has claimed that models do better if you threaten them.
What’s the truth - whilst some of the models seem to have a penchant for sycophancy, does returning this behaviour pay off ?

• A set of three studies from Wharton’s Generative AI Labs has been looking into whether styles of prompting make a difference to the results that a GPT returns. The first study looked at how the question was posed - a formatted prompt, an unformatted prompt, a polite prompt or a less polite prompt. The second study looked at ‘chain of thought (CoT)’ and the third study looked at whether a GPT could be threatened or co-oerced or bribed into better performance.

• The first study unveiled that quality of prompt mattered - but the way the question was asked (polite or impolite) made no difference, as long as the prompt was clear. The second study showed that a CoT prompt is useful for boosting performance but weighed against the increased response times not really beneficial.

• The 3rd study ‘I'll pay you or I'll kill you - but will you care?” is possibly the most entertaining academic paper you’ll ever read (anyone who reads academic papers will likely attest that they tend to be a little dry and devoid of humour)

• This study tried a number of techniques to affect performance, from threatening to punch the AI and even threatening to kick a puppy through to offering tips ranging from $1,000 to $1 trillion for better performance.

• The study showed that threats and coercion had no impact on model performance. Link to the third paper is here, the others can be found in the same vicinity : https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5375404 

Is yours on the list? Top 40 jobs most at risk to AI: Microsoft report

A new report from Microsoft lists the jobs that have the highest cross over with Generative AI and by deduction, are most at risk.
Amongst the most affected occupations are interpreters/translators, historians, passenger attendants), sales reps and writers. Even models and teachers make the top 40

As many parents wonder about the value of tertiary education in an AI future world, the researchers point out that having a degree is no longer a guaranteed path. They do note however that there is “higher applicability [for AI} for occupations requiring a Bachelor’s degree than occupations with lower requirements’

It must be remembered, that much like jobs in the workplace, we are at the early stages of understanding applicability of AI. That specific jobs will no longer be required due to pure crossover with AI capabilities is not a sure fire way of knowing that the job in question will disappear any time soon.

For those looking for a career path that will be potentially less impacted by AI in the future, there’s some good news in the report that it also offers the jobs with the least chance of being impacted by AI - including dredge operators, bridge and lock tenders and water and treatment plant operators.

🔥 In Case You Missed It…

A film is recreated with an alternative AI ending - without director’s knowledge

• Raanjhanaa - a 2013 film about a doomed romance is being re-released with a happier ending. No one told the director, who is somewhat upset

• Released this month, this comes at the same time as rumours of Disney+ planning for users to be able to control more content and endings themselves with AI

• There are now a whole gamut of AI alternative endings that social media users are sharing - such as this creative ending one TikToker shared: https://www.tiktok.com/@batmanandjokervlog/video/7530702608437136670

@batmanandjokervlog

Alternate Titanic Ending? #Titanic #filmtok #veo3 #ai #alternatereality #jackandrose #ending #new

🏆 Tools, podcasts, products or toys we’re playing with this week

Did You Know? The ‘save’ item we use on our computers is a floppy disk, which many GenZ users have never seen. This icon has been around for nearly four decades after being popularised by Microsoft Windows and Apple Macintosh. Well known by all as the thing you press to save, but ask a GenZer to explain what this icon is based on and they may be stumped !

Till next time,

And The Rest Is Leadership