Should we verify EVERYTHING or trust until something is proven false?

The cold war concept of 'Trust but Verify', I agree with, JGunn. Works well on big issues and decisions. The problem today is that we encounter an ENDLESS STREAM of possibly fake images and stories. Personally, I don't do the social media thing, or doom-scroll much, so my exposure is limited, but even the relatively small amount of internet slop I experience is WAY too much to fully verify. So I am left with my own common sense.

If something seems extreme, I wait for multiple sources to agree. Sometimes I'll even research if the issue will significantly change my behavior or world view. But really...how do we trust ANY source of information? Most of them are influenced by a pretty small number of entities that may or may not be interested in actual truth at any given time. So I am ultimately left with common sense again.

What would be best would be a general public that knew the difference between 'I bet that's true' and 'FACT'. Hard reach now that 'Alternative Facts' are an accepted thing for much of the population.

Critical thinking is the only real defense against fakery, and our system has to establish that as Bedrock for education. Whatever that may cost, I'd be happy to help pay.

Depends on what's at stake. Personally, having time on my hands (I'm a retired economist and statistician) I double-check the social media posts of my friends and family, and let them know whether I see verification from a reliable source (e.g., a news organization with a long and consistent track record of fact checking and publication of any errors in its reporting). Mostly that's just an exercise to remind people they should verify before they re-share A.I. slop, so isn't strictly necessary.

But court cases, legal contracts, anything with financial or health consequences - everyone needs to "trust but verify" with the emphasis on "verify." Unfortunately we all have a tendency for confirmation bias and it takes concerted effort to override it until it becomes a habit.

I think the days of trusting something because it is on the web or in print are long over. With AI-generated fake videos looking so real, we need AI to tell us whether they are fake. I think it is a good policy to check the origin of any post or audio, or video on the web with trusted tools.

Verify everything. Society can't survive without agreement on reality. Our disagreements over reality are already tearing society apart, and this trend will get worse as deep fakes improve and become easier to create. Everyday individuals should not be required to pay for this, but platforms and device makers should. If the government is using a potential deepfake artifact as evidence, it should bear the cost of establishing its validity.

That old saying about "Your right to swing your fist ends at the tip of my nose" may have a legitimate application as a constraint on freedom of speech. Freedom of speech must be understood to apply to expressing an opinion, not promoting "one's own alternate facts." This suggests we need to redefine what we mean by a Free Press. But the freedom to lie has never been at the heart of our esteem of multiple reported versions of an event. Publishing as "factual" a manufactured image creates a situation in which the viewer of the image is compelled -- in effect, is made accountable for -- determining the veracity of the information presented. This is analogous to the burden I experience if your fist strikes my nose.

Step one should be a Manhattan Project funded by Government to create very highly reliable watermarking technology -- by whatever actual name -- that either survives cropping, re-imaging, and manipulation, or makes clear that such adulteration of an images has been attempted. As a public good, this technology (likely largely software based?) must be made freely available to all manufacturers of imaging systems and publishers.

Step two would be to hold ALL publicly active platforms liable for failure to apply such technology to any materials they distribute, to determine its provenance, and unambiguously label with a human-readable marker as "Non-AI", or "Not-to-be-taken-as-factual" (or some similarly dichotomous distinction) every image they make available. Any image lacking unimpeachable provenance as unadulterated must be labled non-factual. This labeling need not be obtrusive, or in-your-face, but might be as simple as a single pixel border all around an image in a signal color: perhaps red for AI, green for Factual. The goal wouldn't be to eliminate AI generated or manipulated images, but to make plain to consumers which visuals may be trusted, and which are merely suggestive or entertaining.

Extended visuals, such as videos as compared to still images, would be treated as a whole. A single manipulated frame within a video should result in the entire production being flagged. If future reproduction should exclude the offending changes, perhaps by editing for length only, that edited visual would be evaluated on its own content, not the 'parent' product's.

Publishers and platforms found to be distributing any visual either lacking such a label, or materials mislabeled, must face some drastic consequence, perhaps something like discontinue service for 30 days on a first offence; 90 days on a second, and fully shut down upon a third.

The current branding of YouTube videos as "Includes AI content" is too ambiguous to be meaningful.

The down side of all this of course, is that authoritarian actors like the current administration would have to be subject to an unquestionable burden of proof lest they behave like Big Brother and move against 'enemies' on false pretenses.

Surely the DOJ would exercise that role impartially!

Unlike so many previous new tech introductions which had little or zero forethought into downsides, I think it's imperative we get ahead of this inevitable degradation for once. If the capacity to pass off deepfakes gets solidly entrenched, we will never get control or rid of it.

Just for context, I was part of and present in such new techs starting with the Macintosh in 1983-4. Much benefit, of course, and also much degradation. Especially when algorithms became profit tools in social media. Cf. the movie: the Corporation.

Already we [should] judge the validity of statements-of-fact with reference to the source as we understand it -- familiar news sources develop and maintain a reputation (based on their past history) for reliability and bias; unknown characters on social media don't have such a historical reputation, so we must skeptically weigh the new evidence based on our prior understanding of the world (and everyone should learn to do this!). In the judicial realm, we would do better to further develop (and rely on) logical scientific principles; but that seems to depend on citizens -- in this case judges and juries -- appreciating such a perspective.

I agree with verifying everything. It is alarming to me how many people watch these deepfakes and believe them.

With the avalanche of digital information available from minute to minute, it's essential to differentiate between what's authentic and bogus.

Personally, my "BS detector" is never turned off anymore.

Until something of an inflammatory nature is reliably verified, take it with a "grain of salt".

Adding the feed issue, that being the giant algorithmic driven social media and ad sites that assure we receive ads and content based on history, and the deep fakes and false news become a cocoon of deceit woven by a web of liars to sway opinion. I know this isn't new to any of you, but the reality is humans' capacity to think for themselves seems to be diminishing. And the absorption of this data by everyone, young and old, just adds to my theory that a new mega level of mind control has taken over a segment of our society. When people blindly follow their feed, in spite of the truth, our societies and cultures around the world are in serious trouble. The final blow is the instantaneous and constant barrage of this information. Rather than taking a moment to find the truth we simply move on to the next viral thought pushed by this machine.

Forensics won't matter much in a world driven by dollars and deceit. That is already painfully obvious by simply looking at sites like Snopes that try to provide fact checks. They can't begin to keep up. The complexity of this issue is a real Pandor's box of such magnitude that we likely will never reign it in at this point.

Unlike so many previous new tech introductions which had little or zero forethought into downsides, I think it's imperative we get ahead of this inevitable degradation for once. If the capacity to pass off deepfakes gets solidly entrenched, we will never get control or rid of it.

Just for context, I was part of and present in such new techs starting with the Macintosh in 1983-4. Much benefit, of course, and also much degradation. Especially when algorithms became profit tools in social media. Cf. the movie: the Corporation.

Verify everything !!!!!

It's a scary world when we can't trust our eyes.

we have to find a practical balance between verifying everything and waiting until someone else proves fakery. abraham lincoln said: " you can fool some of the people all of the time and all of the people some of the time but you cannot fool all of the people all of the time:." so we need to learn how to ascertain if something is a fake. we have to have laws on the books, because faked videos or altered videos are a form of libel or slander, depending on how they are used. we have to punish people who do things like this. we have to instill in our children that libel and slander are WRONG. tweaked videos are NOT FUNNY. we need to start thinking about where the wonder of AI will lead us. do we have the ability to pull AI's plug. petroleum engineering is a marvel but we can see some of the problems that have arisen from it. we can see where unregulated stem cell research might lead if we are not careful: the same place as anthrax, which naturally occurs in soil wherever bovines graze, which has been strengthened into a monstrous bioweapon. we need to think about what we are doing and where this might lead us. we need to be able to look at stuff and decide whether it might be true or false: an activity which might be referred to as "critical thinking".

so our default should be a form of skepticism, just not extreme skepticism.

All our perceptions are shaped by our assumptions and biases. Our brains don’t just filter out irrelevant information — they also ignore what doesn’t fit our expectations.

Recognizing this is, I think, the first step.

From there, I think, in an age of AI and deepfakes, a healthy skepticism toward any published photo or video is warranted,

And always consider the source.

(Note that Trump’s campaign and ICE have already published AI-doctored images.)

And I do think, particularly in the news media, that both in-house and independent agencies dedicated to spotting and calling out AI-generated images is warranted.

It would be optimal if everything could be verified. But that's impossible. I would like all content I receive to be verified. I don't have time to verify it myself, and I'm online to communicate with colleagues or friends and to learn from researchers and professional journalists. In other words, I'm online to learn, not to be played with.

I have no doubt that verifying everything is impossible. Therefore, we should avoid sharing everything we receive unless we take care to verify the source.

Without the availability of methods of truth verification and falsity detection, we cannot trust the internet, and our society falls prey to immoral agents of chaos, destruction and greed. I believe that successful research on verification and falsity versus reality detection is a necessary condition for the survival and advancement of civilization.

At this point I think that EVERYTHING needs to be assessed for truth and validity. Because the web lets anything in and even those who know most truths can still be led astray by falsehoods. We need to flood the internet waves with truths so it can’t lead anyone astray.

I'm not sure it's that complicated. A default position of disbelief should be adopted for videos posted by extremist groups that generally wish to push a specific agenda. Newsmax, Fox News, The GOP, and anything posted on Truth Social, should never be believed unless that video is also featured by other, more credible, organizations. And believe nothing that appears sensational on any social media platform, or anything forwarded by friends from those sources.

Deep fakes are corrupting our sense of reality. There indeed needs to be platforms that verify things posted as facts, because too many people are believing things that are negatively impacting our relationships with others, our government, and our sense of what is right and wrong.

The costs should be borne be the creator of the deepfake. But creators of the technology should post a large bond that makes then liable for use of their technology. On the other hand, evidence that isn't evidence can't be used as evidence.