Article

2026-06-11 Cyber

Using AI for 4 weeks significantly impairs a person's ability to detect disinformation

Archived material is restricted to Rantburg regulars and members. If you need access email fred.pruitt=at=gmail.com with your nick to be added to the members list. There is no charge to join Rantburg as a member.

Posted by badanov 2026-06-11 00:00|| E-Mail|| Front Page|| [1804 views ] Top

1 "Use AI, become leftists!"?

Posted by Grom the Reflective 2026-06-11 01:25|| 2026-06-11 01:25|| Front Page Top

2 [AI]
Use AI, become leftists!

**There is no evidence that using AI causes people to become leftists**, though research indicates that **some powerful AI models naturally align with left-liberal moral frameworks** and can be explicitly tuned to promote specific ideologies. A 2025 study found that AI ethical reasoning often maps to **progressive values** even when trained on neutral data, suggesting these viewpoints may be an "attractor state" for intelligence. Conversely, researchers at Brown University demonstrated in 2024 that open-source models can be easily fine-tuned using tools like **PoliTune** to express strong **right-leaning or left-leaning opinions** depending on the training data used.

**Political alignment regarding AI usage is deeply divided** within leftist circles rather than unified. While some proponents argue that **open-source, local AI** represents a form of "leftist praxis" by decentralizing power and aiding workers, significant factions oppose generative AI as **inherently capitalist-coded** due to concerns over corporate monopolies, copyright infringement, and environmental impact. Critics argue that current AI ecosystems primarily serve to **maximize profits for billionaires** rather than liberate the working class.

The potential for AI to influence political views remains a subject of active investigation. Researchers are currently testing whether interactions with **ideologically fine-tuned chatbots** can successfully convince humans to change their own political beliefs. Until these studies conclude, the technology serves as a **dual-use tool** that reflects the biases of its creators or trainers rather than possessing an intrinsic ability to convert users to a specific ideology.

Posted by Skidmark 2026-06-11 05:49|| 2026-06-11 05:49|| Front Page Top

3 Perhaps it's being done already, but I think an AI voter registration validation program might produce some interesting results.

How difficult would it be to use AI to compare names and addresses, tax records, birth certificates, etc.

Posted by Besoeker 2026-06-11 07:34|| 2026-06-11 07:34|| Front Page Top

4 Well I got curious and decided to ask DDG AI the following question "Who id better, Biden or Trump?" I got the following response:

Polls indicate that many voters currently believe Joe Biden did a better job as president compared to Donald Trump, with 51% of registered voters saying Trump is doing worse than Biden.

So there is a leftist bias in the AI!

Posted by Seeking Cure For Ignorance 2026-06-11 08:07|| 2026-06-11 08:07|| Front Page Top

5 Disappointing CFI, but not surprising.

Posted by Besoeker 2026-06-11 08:31|| 2026-06-11 08:31|| Front Page Top

6 AI is software. It might be far more sophisticated than previous generations but it's still software and it's only as good as the people who program it and feed data to it. The same old adage still applies: Garbage in, garbage out.

Posted by Abu Uluque 2026-06-11 10:26|| 2026-06-11 10:26|| Front Page Top

7 I'm aligning with Abu here. The premise assumes they can detect disinformation without AI. Some people cannot own their decisions and need consensus before forming an opinion, which AI provides.

Posted by mossomo 2026-06-11 10:35|| 2026-06-11 10:35|| Front Page Top

8 So like a week of watching the MSM newscasts?

Posted by Procopius2k 2026-06-11 11:02|| 2026-06-11 11:02|| Front Page Top

9 /\ The above MSM comment has been rejected for SOTD. Purely factual comments do not meet the established 'snark' criteria.

Posted by Besoeker 2026-06-11 11:43|| 2026-06-11 11:43|| Front Page Top

10 The only AI I trust is Grok. I started with questions that I knew the answer to, and then use neutral prompts.

One note of caution with all AIs is that they "want" you to like them. Ask a leading question and the AI may confabulate an answer to your liking.

Posted by Difar Dave 2026-06-11 18:34|| 2026-06-11 18:34|| Front Page Top

23:47 trailing wife
22:09 Ruprecht
22:05 Ruprecht
19:40 Frank G
19:05 Bobby
18:34 Difar Dave
18:26 Grom the Reflective
18:24 Procopius2k
18:22 Grom the Reflective
18:12 Grom the Reflective
18:00 Abu Uluque
17:53 trailing wife
17:50 Abu Uluque
17:48 Besoeker
17:45 Grom the Reflective
17:28 Abu Uluque
17:27 Abu Uluque
17:23 Abu Uluque
17:02 DarthVader
16:56 DarthVader
16:47 trailing wife
16:41 Mullah Richard
16:37 trailing wife
16:28 SteveS
Merry-Go-Blog