# Two‑Thirds of Americans View AI Progress Negatively, New Surveys Show

> A new poll shows 64% of Americans view AI development negatively, while firms report rising ROI and experts warn of urgent risks, setting up a clash between market optimism and public concern.

- **Published**: 2026-09-17 19:30:38
- **Canonical**: https://worldys.news/article/two-thirds-of-americans-view-ai-progress-negatively-new-surveys-show

## Reporting

A fresh Gallup‑style poll released this week found that almost two in three U.S. adults consider the rapid development of artificial intelligence to be a bad thing, marking the sharpest dip in public optimism since the technology first entered mainstream conversation. The finding arrives as corporate leaders tout rising returns on AI investments and experts warn of “urgent” risks that could shape regulation and market behavior through 2026.
Core developments
The Hill reported that 64% of respondents in the survey answered “yes” when asked whether AI development is a bad thing, while only 19% said it is a good thing and the remainder were undecided. The same poll highlighted that concerns are strongest among older adults and those with lower levels of formal education.
At the same time, McKinsey & Company’s 2026 outlook on AI shows that firms that have moved beyond pilot projects are beginning to see measurable returns. The consultancy notes that enterprises that embed AI in core processes are reporting productivity gains ranging from low‑single digits to double‑digit percentages, depending on the sector. However, McKinsey also cautions that the “AI ROI gap” remains wide: many organizations still struggle to translate proof‑of‑concepts into profit‑center outcomes.
Parallel data from Ipsos reinforce the mixed picture. Their global survey of 12,000 adults found that while 71% have interacted with AI‑powered chatbots or voice assistants in the past year, only 38% feel that AI makes their lives easier overall. The Ipsos report attributes the gap to perceived loss of control and worries about data privacy.
In a complementary Pew Research Center study, 58% of Americans said they are “somewhat” or “very” concerned about AI’s impact on jobs, and 44% expressed unease about AI‑generated content influencing public opinion. The Pew survey also revealed that 62% own a smart device (such as a speaker or thermostat) that uses AI, yet a majority still lack confidence in the technology’s long‑term societal effects.
From an expert‑risk perspective, MIT Sloan’s analysis of 272 AI specialists identified “autonomous weapons,” “deep‑fakes and misinformation,” and “uncontrolled self‑improving systems” as the top three urgent risks. The report stresses that these risks are seen as more immediate than longer‑term existential concerns.
Legal professionals surveyed by Thomson Reuters echoed the risk theme, noting that AI’s role in contract analysis and litigation support raises questions about accountability, bias, and the adequacy of existing legal frameworks. Over half of the respondents said their firms are currently drafting AI‑specific policies, but only 22% believe current regulations are sufficient.
Why it matters
Public sentiment matters because it can shape policy and investment. When a majority of citizens view AI development as detrimental, legislators are more likely to introduce stricter oversight, as seen in recent bills targeting facial‑recognition technology and algorithmic transparency. Such regulation could slow the rollout of AI‑driven products, affecting the timelines projected by firms seeking ROI.
Conversely, the McKinsey data suggest that businesses are already reaping financial benefits, creating a tension between market forces and public opinion. Companies that ignore consumer wariness risk brand damage, especially in consumer‑facing sectors like retail and media where trust is paramount.
The risk rankings from MIT Sloan and the legal community underline the potential for AI to generate harms that exceed ordinary business losses. Autonomous weapons and deep‑fakes, for example, could trigger geopolitical instability or erode democratic processes, prompting governments to act decisively.
Finally, the disparity between adoption rates and confidence levels—high usage of smart devices paired with low trust—signals a gap that could hinder the next wave of AI‑enabled services, from personalized healthcare to autonomous transport.
What the sources show
All six sources agree that AI is both increasingly embedded in daily life and the subject of growing unease. The Hill’s poll provides the most striking headline figure—64% negative perception—while Pew adds nuance by breaking down concerns about employment and misinformation. Ipsos contributes a usage metric, showing that a large majority have interacted with AI tools, yet a smaller share feel positively about those interactions.
On the business side, McKinsey’s report is the only source that quantifies financial outcomes, noting that firms achieving “operational AI” are beginning to see tangible productivity lifts. However, McKinsey also points out that many companies remain stuck in the “pilot” phase, a gap that aligns with the public’s skepticism about AI’s real‑world value.
Risk assessments differ in emphasis. MIT Sloan’s expert panel prioritizes weaponization and misinformation, whereas the Thomson Reuters legal survey focuses on accountability and regulatory adequacy. Both converge on the view that current governance structures are insufficient.
Where the sources diverge is in the projected timeline for widespread AI benefits. McKinsey forecasts that by 2028, at least 30% of global GDP could be linked to AI‑driven efficiencies, while the public‑opinion surveys suggest that acceptance may lag behind that economic trajectory.
What’s next
Legislators are expected to introduce at least three new AI‑related bills in the U.S. Congress by early 2027, targeting algorithmic transparency and high‑risk applications, according to a policy tracker cited by The Hill. Companies that have already integrated AI into revenue‑critical processes are likely to double down on investment, aiming to meet the ROI benchmarks outlined by McKinsey before the end of 2026.
Experts from MIT Sloan recommend that firms adopt “risk‑first” governance frameworks within the next 12 months to mitigate the top three threats they identified. Legal professionals surveyed by Thomson Reuters say their firms will continue to refine AI policies, with a target of full compliance with emerging standards by the second quarter of 2027.
For the broader public, Pew’s longitudinal study will release a follow‑up in 2028 to track whether attitudes shift as AI becomes more visible in education, healthcare, and public services. The next wave of consumer sentiment data will be crucial for aligning regulatory action with societal expectations.

---
*Synthesized by Worldys News Intelligence Desk under journalistic verification standards.*