52% Use AI, Meaning of Manus, AI.Gov, Job Shifts

Thinking Out Loud

52% of Americans Use AI: When Dialog did our first consumer survey there were perhaps one or two others in the market. Now there are at least a dozen. The latest comes from Elon University’s Imagining the Digital Future Center. The survey of 500 users in January extrapolated that 52% of the US population has used/uses AI tools. The most popular is ChatGPT (72%) followed by Gemini (50%). Dialog's earlier survey found that 44% of US adults had ever used AI in Q4 last year. The top five Elon survey use cases for AI were: searching for facts quickly like a search engine (68%), researching new subjects/ideas (64%), brainstorming about ideas (55%), summarizing and understanding written material like documents or emails (47%), and creating emails (42%). Personal uses (51%) predominate over work use cases (24%). The majority (76%) report being generally satisfied with results. While Google argues it hasn't been affected by AI, these consumer surveys strongly argue that Google is vulnerable to displacement for informational queries and consumer research – for which AI is much better suited.

Future of (Local) News?: The Arizona Supreme Court has introduced two "AI reporters." These are synthetic images and voices that are real enough to be somewhat convincing. They will reportedly offer updates on court rulings, legal developments, and public services and will be available on the court's website and social media accounts. Japan has had AI news readers for some time. But this development suggests it will start to become more common in North America. You can imagine all kinds of online news publishers adopting this approach to boost online engagement and, eventually, even for traditional news broadcasts. It's obviously cheaper than humans. And as the technology becomes increasingly capable and lifelike, it may be equally acceptable to audiences – even preferred.

Better Management via Bot: Most people who quit their jobs do so because of a bad relationship with their boss. There are huge costs and lost productivity associated with turnover. Yet people who get promoted into management roles are usually good at their jobs but don't necessarily know how to manage people. They receive little or no training and often don't have the "soft skills" to be effective managers. Enter AI management coaches, which can potentially help new or mediocre managers develop better skills in dealing with direct reports. While this doesn't replace quality human management training, in a world where most companies don't invest in training or place much emphasis on soft skills (vs. KPIs) it's better than nothing.

Struggling SMBs: I recently co-led an SMB marketing workshop in San Francisco. Most of the attendees were relatively new service businesses. There were also established small companies and consultants seeking to learn more about digital marketing. The sophistication of attendees ran from low to relatively high. Consistent with what we found in our first SMB survey last year, they were almost all on social media. Most also had websites. But few had claimed the Google Business Profiles (some had). Our initial survey showed only 44% were managing business profiles on Google or Yelp. The vast majority of people at the workshop were using ChatGPT. SMBs will gravitate toward easy to use, high-value tools. In contrast, verifying your business on Google can be challenging and frustrating. They also need trustworthy marketing advice, which they're not really getting. Many agencies and vendors are still mostly pitching while calling it "education."

News & Noteworthy

• Top 50 web and mobile AI apps: ChatGPT leads both; Gemini #22.
• Claude has become the coding agent of choice for enterprise developers.
• Why the near-term economic impact of AI may be less than expected.
• Musk's Grok outperforms ChatGPT on certain benchmark tests.
• Is OpenAI's latest release a dud? Some critics think so.
• Auction house Christie's AI art auction brings in more than expected.
• Amazon and MSFT mostly exiting the hardware/gadget business.
• Spain might fine up to 7% of all revenue for failing to label AI content.
• Majority of CEOs think AI would outperform human board members.
• Stanford study documents AI content proliferation across the web.
• AI may force law firm business model change: the end of "billable hours."
• Allstate: AI doing a better job than reps, communicating with customers.

Meaning of Manus

Manus is being promoted as the "first general AI agent." Created by Chinese startup Butterfly Effect, Manus appears more advanced than other US-based AI agents (e.g., OpenAI's Operator). The demo shows three different tasks: screening resumes, looking for NYC rental property and stock analysis. There are many more use cases on its website. Like DeepSeek, Manus was immediately heralded a formidable challenger to OpenAI and others. But there are skeptics, and users have reported mixed results. Deep Research from OpenAI and Gemini (same product name) are also agents that can replace a considerable amount of manual effort online. The improvement and ascendancy of agents will ultimately eat into traditional search and change the way people plan, discover and buy things. For example, think about SaaS tool comparisons and how agents could dramatically reduce the time involved in getting to a buying decision – while the buyer never directly visits Capterra, G2 or does any direct "web searching." The AI does the informational and consideration steps for you, with major implications for B2C and B2B marketers.

Government by AI

Whatever you think of the (indiscriminate and reckless) dismantling of the US federal government by DOGE, there's a massive Silicon Valley style experiment going on. Musk and his minions seek to replace human employees with AI agents on an unprecedented scale. As The Atlantic reported, "The idea is simple: use generative AI to automate work that was previously done by people." Chatbots have already been deployed and are being tested at the General Services Administration. While there's nothing wrong with trying to judiciously utilize AI to boost efficiency or improve outcomes, the mass replacement of people with AI carries very real risks for cybersecurity, national security, and the delivery of services. AI systems are vulnerable to hacking and surely will be hacked by cybercriminals and foreign adversaries (China et al). AI bias and automation errors could disrupt benefits, wrongly deny services and erode public trust. Also, the majority of Americans don’t want government to make decisions with AI. We'll see what happens.

AI Will Transform Job Market

AI adoption has been uneven to date. But its impact on jobs is starting to emerge. According to the World Economic Forum's (WEF's) Future of Jobs Report 2025, AI is expected to create millions of new jobs but reduce demand for others. In one sense this is true of all new technologies. But AI's impact is likely to be more sweeping. WEF has projected the fastest growing and fastest declining jobs. These predictions are not based solely on AI but it's a factor. Lower level white collar, clerical and administrative jobs with repetitive functions will reportedly see the biggest losses; also: cashiers, telemarketers, accounting clerks, payroll clerks and bank tellers. The biggest gains will be in fields built around data, software and AI: "big data" specialists, fintech engineers, AI/machine learning experts and software developers. These predictions are partly based on a survey of "over 1,000 global employers" from numerous countries, as well other data inputs. We should see it only as directional. In some cases it could be entirely off the mark. Software/app developers is near the top of the high-growth list. But this category is especially vulnerable to AI displacement. The WEF ultimately projects that 92 million jobs will be eliminated but 170 million new jobs will be created, resulting in net job growth on a global basis.

Funny | Disturbing | Sad

• Because of labor shortages, cat-themed robots waiting tables in Japan.
• Humanoid robots are advancing very quickly – perhaps too quickly.
• Russia has infected major chatbots with anti-Ukraine propaganda.
• AI bots are regularly crawling publisher sites but not driving enough clicks.
• ChatGPT has unleashed an "era of cheap speech."
• Study: AI infrastructure has potentially serious public health consequences.
• More and more lonely people turning to AI bots for companionship.

If you're not a member and would like to explore Dialog membership, contact us. In 2025, we have expanded benefits but will limit new members.

If you're not getting this in your in-box, you know what to do...