How to tell if you’re talking to an online fake

People fake social media profiles for many reasons, but none are good. Technology may change and make setting up false social media accounts easier, but the reasons for trying to dupe others have existed for centuries.

Chatbots and fake social media accounts are about following you into believing you are interacting with a human but why the deceit?

Here. We look at how chatbots and online fakers work and what they try to achieve.

How many fake accounts are there?

Surveys reckon that one in three social media users have a fake account, and half have two or more pretend identities.

Their motives are many and varied.

To share thoughts without being judged garnered the most support, with 41 per cent of the vote in research.

Spying on someone else’s profile was the second most popular response from 38 per cent of users.

Others were more honest in smaller numbers – 13 per cent thought a fake account would help them win online contests, 5 per cent manipulated a fake account to pump other users to follow their real social media account, and 0.4 per cent confessed they were crooks out to scan others.

And where are these fake accounts?

Twitter is the most infested, with 41 per cent of users thought to be faked. Facebook (31%) and Instagram (28%) comfortably grab second and third rankings.

Facebook has 3 billion active users each month, meaning around a billion user accounts are bogus. Instagram has about 1.5 billion active users, making 420 million fakes. Twitter has 330 million users, of whom 135 million don’t exist.

Assuming many social media users are active across multiple networks, web monitor Search Logistics numbers suggest 3.5 billion of us use social media regularly. If that’s the case that 0.4 per cent of scammers represent at least 14 million crooks.

And who do the fakers pretend they are?

Chameleon crooks shapeshift their identities to match the reason for their duplicity.

An overwhelming 80 per cent pick a same-sex profile, while 13 per cent swap genders. Around 3.5 per cent pretend to run a business, and 2.5 per cent masquerade as celebrities. Many of these fake accounts are operated by chatbots.

A chatbot is an automated software application that mimics human conversation. Chatbots have artificial intelligence designed to communicate with someone by simulating human actions.

One of the most popular chatbot applications is customer service or virtual assistance.

How to tell if you’re talking to a bot

Artificial intelligence is becoming so sophisticated it’s a challenge to work out if you are talking to a real person or a bot.

The most common way to detect a fake is to check the bot’s social media profile. Most bots neglect to place photos, links or personal information on their bio page. Some can upload an image, but fakes can be detected with Google’s reverse image search or TinEye, another search program.

Bots also have problems duplicating the nuances of human language. Bots miss jokes and irony and repeat responses or suddenly change track.

Because bots are designed with a single objective, they tend to have one-track machine minds that repeatedly deliver the same comments or links.

Timing is another giveaway. Few people post social media comments round-the-clock. Review the profile posts to see if you can spot a pattern. Most people have routines and fit their posts around their downtime. If a bot is posting continually, then it’s probably a fake.

Friends and followers can be a giveaway. Bot accounts tend to have few social relationships with other users.

Why don’t social media sites crack down on fakes?

Social media sites do not delete fake accounts because most do not break community rules or cause a nuisance.

Providing the fake account is run by an individual; the faker is still seeing advertising and not firing off thousands of interactions like a bot. Advertisers and social media sites do not track advertising views, not which accounts view them, so everything works as it should on the business side.

How do chatbots work?

Humans are predictable and generally behave to a pattern. Looking back over customer service data will likely show people call a business for limited reasons – for support, to order or return a product, to pay or query an invoice.

In the past, a simple bot had predetermined responses triggered by keywords that worked like an online FAQ.

A giveaway you are communicating with a learning AI bot is if you ask a question, and get a reply, then the bot moves on to ask a follow-up question.

For example, ask a bot if it will rain tomorrow. A simple bot will answer something like yes, at around 9 am. An AI will give the same answer and then ask if you want to get up earlier to avoid a longer commute. AI bots learn to link information like humans.

And any communication you have online is more likely to be with a chatbot than with a human.

Wikipedia lists almost 30 chatbots in general use across the web by millions of sites. This shows that most of us have talked with a chatbot without knowing.

The Future of Chatbots

Chatbots are artificial intelligence taking over the humdrum tasks humans don’t want to do. They aren’t going away anytime soon, but some developers want to push the boundaries.

The line to cross is will we know we are chatting to a bot, and should a bot tell us it’s an AI from the outset of any dealings?

At this time, AI is far from perfect but refines and grows with every interaction.

English is especially tough for AI to handle, as the language is an illogical mish-mash of mispronounced words and strange grammar rules. This makes chatbots and AI a little clunky and easier to detect.

As the sector matures, politicians say we should regulate bots. British Prime Minister Rishi Sunak wants Britain to head up the lawmakers and is seeking support from the US to expand the UK research that is already taking place.

Chatbot and AI FAQ

Here is an FAQ about AI and chatbots written by a chatbot chatGPT