Again, lies and deception are nothing new. However, the efficiency and presumed exactness of AI will make most people less likely to double check, which will make it easier for dishonest actors to deceive a larger number of people.
Artificial intelligence is efficient to a degree, but it is ultimately limited.
Even when there is no dishonest or under-handed intent, we all have to be aware of content we encounter, not least the internet, although to say so is a bit of a cliche nowadays. An example is Wikipedia which purports to be the font of all knowledge, but anyone with any awareness and knowledge will know that the information on that website cannot be trusted. Even the science-based articles on that site have to be treated with caution. Not just Wikipedia, but also well respected scientific journals - this one https://www.ncbi.nlm.nih.gov/pmc/articles/PMC61045/
beggars belief! To give BMJ a little credit, they published https://www.ncbi.nlm.nih.gov/pmc/articles/PMC139029/
that refuted the claims. Other subject matter that attracts less scrutiny, such as history and personalities, are much less reliable. Of course, artificial intelligence isn't responsible for the content, rather it is the ill-informed, or individuals intent on deception. In this regard, artificial intelligence is superfluous and inherently naive.
One interesting aspect is the use, for many years, of artificial intelligence in the scientific community to show up severe short-comings in the publishing of scientific research. To show the lack of peer review, loose standards, (and the placement of monetary interest over science [my additional reason]), Alan Sokal submitted, way back in the 1990's, a non-sensical article to a journal that was subsequently published (and then retracted by the journal once Sokal revealed his hoax). You can 'google' it. On the back of this hoax/affair, a number of people (usually students) have since used artificial intelligence to create scientific articles and submit them to not only dubious publishers, but also some well known ones where the article went throught the peer-review process. Many have been published. These articles often contain all the right words in all the right order, but make no sense whatsoever, i.e., gobbledygook. These examples were used to highlight the loose standards shown by a number or publishers (and reviewers who should have known better). Suffice to say that science has been well aware of artificial intelligence for some time now, and it's also aware of its innate limitations. A scientific article created by artificial intelligence, with its included introduction, methodology, results, discussion, etc. will lack one very human factor. Reason. That and the inability of independent scientists to duplicate or repeat the experiment/observation/results because the paper was a load of cods...
I should add that this is not a reason to distrust science itself (that would be incredibly ignorant), just those that purport to publish it.
Just to make everyone more depressed, at least for those that don't know, there are now a number of websites that allow you to create your own computer-generated article. By inputting keywords, the websites will generate your own research article, poem, song lyrics, etc. However, these written forms tend to be either non-sensical, ambiguous, or contain content it garnered from other websites such as Wikipedia, or combine all three aspects. You can try it yourself. Just search for 'ai paper generator'. Add some swear words to the keywords for an interesting result!
Of course, while science can quickly detect articifiial intelligence (intelligence being a dubious moniker), in areas where it stands up to less rigour such as the media, social culture, internet, it can be less easy to detect, namely, as Jorgen points out, due to ignorance. However, artificial intelligence can only have a limited effect as behind all artificial intelligence is a human being manipulating it in accordance to his or her design. In addition, age old lies and deception will always trump artificial intelligence so one can be assured that artificial intelligence cannot maniplate politics (and religion for that matter) to any meaningful degree as that's already the forte of very dubious politicians and media.
Artificial intelligence cannot tell you the answer to life, the Universe, and everything, or if there is a god (with apologies to Frederick Brown on the latter). Like any other computer program, it can only generate a result given a set of inputted parameters. Those parameters are set by human beings who have, in this case, a limited understanding of the Universe. The closest any artificial intelligence can get to answering such a question will be the same as that currently answered by the human being inputting the parameters. It's better suited to adding two and two together on that other form of artificial intelligence (a calculator) where the parameters are fully known and understood. Artificial intelligence is limited by human intelligence. With regard to art, whether it's a poem, song, or photograph, it will lack the underlying human emotion or meaning intended to be relayed or envoked by that poem, song, or image. Lacking a human 'soul', for want of a better description, artificial intelligence is wholly inadequate. As with science without reason, the art of artificial intelligence will lack meaning and intended emotion.
...and if one's art is likened to that of artificial intelligence, then it could replace the old addage of "a monkey could do better than that". ...and I'm pretty sure it could do better.
P.S., Reading back of my waffle of pants, it makes me wonder if that was computer generated too!