Mariawas sitting in front of her computer in the small marketing agency in Sofia, where she worked as a copywriter. Her fingers froze over the keyboard as she stared at the email she had just received, “Sorry, but your services are no longer needed. We’re switching over entirely to AI-generated content.” What Maria didn’t yet know was that this moment would prove to be part of a much larger and troubling transformation…
The new reality of digital marketing
In 2024, artificial intelligence has taken centre stage not only in the news, but in people’s everyday lives, but with it comes the next problem – AI-generated misinformation has been identified as the most serious near-term global threat for the next two years.
According to the World Economic Forum’s‘Global Risks Report 2024‘, misinformation created by artificial intelligence has been identified as the most serious near-term global threat for the next two years. Artificial intelligence systems trained on data from the internet can generate false information that threatens not only democratic processes but also the business environment as a whole.”
Unfortunately, we don’t realize that when we replace human creativity with AI algorithms in marketing, we’re not just optimizing our costs. We’re fundamentally changing the way society perceives and processes information, and it’s hurting not only ourselves, personally, but businesses as well. And the damage done can be many times greater than the investment earned in marketing and advertising. Let me spare you this …
Anatomy of transformation
According to a World Economic Forum report on digital risks (), generative AI is creating unprecedented challenges for the information environment. A specific example is LAION5B, the largest publicly available AI training database, which contains over 5 billion web links to images and texts used to train systems such as Stable Diffusion and DALL-E. This huge database, while a valuable resource for AI development, raises serious questions about the quality and reliability of the information used to train AI systems.
In Bulgaria, the scale of the change is already being felt. In the rush to attract clients at any cost, marketers, freelancers and agencies, en masse, started to lower prices and replace their services with artificial intelligence – from the first ever marketing strategies created by artificial intelligence (for 200 BGN), to copywriting AI services for 50 to 500 BGN / month, automated video production for 200 to 2000 BGN for video and visual content generation: from 30 to 300 BGN for a package of images.
Unfortunately, no one is yet talking about (or prefers to gloss over) the fact that artificial intelligence can be easily manipulated and is still unable to distinguish between proven and false information, data sources and ideas. This means that a so-called ‘marketing or business strategy’ generated with AI may not only be the product of complete fabrication, but it is also likely to be close to the strategy of a competing industry in the same niche.
“In their quest for rapid digitization, companies risk feeding their AI systems with low-quality and manipulated data, which could lead to the multiplication of errors and misinformation in the future.”
The social consequences of the era of misinformation
The implications of this development are multi-layered and profound. A Bloomberg Digital News Report 2024 study shows that 78% of Americans get their news primarily from social media, where algorithms create so-called “information bubbles” that isolate users from alternative viewpoints. Even more alarming, 45% admit they rarely check the sources of the information they share.
“The problem of misinformation in AI systems creates a vicious cycle for businesses. According again to the World Economic Forum research I cited above, AI systems trained on data from the internet often fall into the trap of so-called ‘algorithmic hallucinations’ – a phenomenon where systems prioritise plausibility over factual accuracy of information.
For businesses, this creates a double risk: on the one hand, marketing strategies and business decisions based on AI analysis of poor quality or manipulated data can lead to wrong decisions. On the other hand, public trust in digital content drops dramatically, directly impacting the effectiveness of marketing efforts.
Furthermore, when people stop trusting verified sources of information, they become more susceptible to manipulation and conspiracy theories.
Of particular concern is the fact, documented in the report“The Risks of Generative AI,” that AI systems can inadvertently multiply errors in market analysis and consumer behavior, creating a distorted view of real customer needs and preferences that can be measured in concrete losses to the business.
At the moment, governments are struggling to push through sensible frameworks and laws to manage this technology and its drawbacks, meaning companies have to tackle this problem themselves.
Harvard Business Review classifies the risks and challenges of integrating AI into business as four groups of risks.
Misuse of Artificial Intelligence (AI)
Misuse refers to the unethical or illegal use of generative AI capabilities for harmful purposes, e.g. fraud, cybercrime and disinformation campaigns.
The reduced cost of creating content with generative artificial intelligence has led to an increase in so-called deep-fake content, which is used for social engineering, automated disinformation attacks, fraud, identity theft, and manipulation of election results.
Improper application of artificial intelligence
Artificial intelligence “hallucinations” can lead to distortions of information and facts. In June 2023, a radio host from the U.S. state of Georgia filed a lawsuit against OpenAI after ChatGPT incorrectly claimed he was accused of fraud and misappropriation of funds by a non-profit organization.
Misrepresentation of information by artificial intelligence
This category of risks covers scenarios in which artificial intelligence is used purposefully to spread false information, such as the photos and videos of Tesla’s truck crashes that are circulating on the web.
Manipulating financial markets for public companies
The marketing of public companies and IPOs is a topic that excites me vividly, so somehow I can not share that artificial intelligence and the false information it spreads can directly affect the stock prices of public companies. Which is a problem. Especially if the company doesn’t have a marketing team and doesn’t care about its public image and communication with investors.
An example of this is the Pentagon explosion publications that pummeled Wall Street stocks because many people believed it actually happened.
And more …
What could be the risks for your business if your AI-generated marketing strategy turns out to be… umm, not quite adequate
Develop a marketing strategy based on AI analysis – there is a risk that it is based on fake reviews and manipulated reviews, leading to:
- 60% drop in marketing campaign effectiveness.
- Up to 300% increase in customer acquisition costs.
- On average, 43% of the product development budget goes to features that don’t meet real market needs.
Creating an Ideal Customer profile based on information and data from bots and fake profiles:
- Between 15% and 30% of social media profiles show signs of automated or manipulative activity.
- On average, companies invest around 10-25% of their annual turnover on marketing strategies – imagine the losses if they are based on unreliable data.
Making decisions to enter new markets based on AI analysis:
- An average 54% drop in consumer trust in companies using AI-generated content without referencing data or citing data and sources.
Misinformation, AI and rethinking business models
Yesterday, 07.01.2025, in his personal profile, Mark Zuckerberg made a historic address and said, “It’s time to go back to our roots – a platform where communities themselves define what is true. I believe in the collective wisdom of our users more than any centralized fact-checking system.”
Facebook reaches 2.9 billion monthly active users. What Mark doesn’t talk about in his statement, but should, is fake profiles.
It is speculated that the number of fake Facebook profiles is currently between 15% and 30% (and possibly more) or as reports suggest – “between 15% and 30% of profiles show signs of automated or manipulative activity“.
What does this mean?
First of all – Mark has sold his soul to the devil and it is now quite official that Meta’s existence depends on whether and how it serves Trump’s interests – it is no coincidence that Zuckerberg moved all content moderation from California to Texas, where Trump won the Texas election for the third time in a row.
But the second implication is that social networks will build an increasing role in spreading misinformation and feeding artificial intelligence into fake data.
Elon Musk with X (formerly Twitter) plays a role in this.
Elon Musk is not just one of the richest people on the planet – he is a visionary, a role model for many, but he is already on his way to becoming one of the most influential people on the planet and directly influencing global political processes.
- Elon Musk supports the AK Party and criticises German President Frank-Walter Steinmeier.
- Vice Chancellor Robert Habeck described Musk’s comments as an attack on the EU aimed at weakening it.
- The multi-billionaire also intervened in UK politics.
These are just some of the headlines his name has been associated with in the last few weeks since Trump’s victory. But before that, of course, Musk announced a public endorsement of Trump’s presidential campaign, accompanied by a $5 million donation and free ad space on X worth more than $4 million.
And if you think that this does not affect us in Bulgaria – nothing like that.
Still small and medium with budgets up to 1000 BGN/month rely mainly on Facebook and Instagram to sell online and advertise their products and services.
Now imagine how that 1000lv. goes to train the AI bots and target fake profiles. Doesn’t look encouraging, does it?
Wait, it could get worse.
Imagine a world where you can’t believe anything you see or read. A world where truth is like sand between our toes – the more we try to hold on to it, the faster it slips away. Misinformation created by artificial intelligence is the most serious near-term global threat for the next two years. And this is not just another alarmist forecast – it is our reality today.
The Sunset of Democratic Values: From Washington to Beijing
In recent years, we have witnessed an unprecedented offensive against democratic foundations worldwide. The rise of authoritarian leaders such as Donald Trump in the US, Vladimir Putin in Russia, Xi Jinping in China, Recep Tayyip Erdogan in Turkey, Aleksandar Vucic in Serbia and Viktor Orban in Hungary marks a worrying trend – the gradual retreat of democratic values before populism and authoritarianism.
“Over the past two years, the Committee on Foreign Interference has gathered comprehensive and hard evidence of foreign interference and manipulation of information online in numerous liberal democracies around the world by malicious forces,” Sandra Kalniete (EPP, Latvia) told the European Parliament. She authored the Parliament’s report, which asks the Commission to propose a multi-layered strategy of measures to deter hybrid threats and attacks from other countries. The strategy should include collective sanctions, measures to improve media literacy, counter social media campaigns and the secret funding of political parties, strengthen cyber security and the protection of EU institutions.”
Artificial intelligence: the new weapon of mass misinformation for politicians
As artificial intelligence grows in popularity, so does its role in misinforming the public. Misinformation that authoritarian political leaders support, apparently with the backing of major social networks and media, as part of their propaganda and “entrenchment” strategy in power.
And lest my words sound like a “conspiracy”, let us look into the biography of just one of them – Aleksandar Vucic. The pattern and experience of the others is no different.
Aleksandar Vučić, the current president of Serbia, who rules the country authoritarian, although Serbia is a parliamentary republic, was minister of information (read “propaganda”) and a member of the leadership of the Serbian Radical Party of convicted war criminal Vojislav Šešelj, whose paramilitary groups have committed countless war crimes in Croatia and Bosnia.
It is no coincidence that freedom of media and propaganda is one of the biggest problems in Serbia at the moment. If you have spent more than a week in Serbia and follow foreign policy, you will know that this is not just ‘idle talk’ and ‘speculation’.
“Serbia has been negotiating for EU accession for ten years, but at the same time maintains close relations with Russia and considers itself its younger Balkan brother. Authoritarian ruling President Aleksandar Vucic often emphasises Moscow’s role as Belgrade’s patron and has repeatedly received Russian President Putin with great honours.”
Deustsche WElle
One last note on the political topic before I continue with why blind trust in artificial intelligence (AI) can be disastrous for business – a Deutsche Welle report on trolls, fake profiles and how misinformation works.
“Everybody lies – how the internet reveals the secrets of human nature”
This is one of my favourite books by Seth Stevens-Davidovitz, which although published 6 years ago by Ciela in Bulgarian remains a title I recommend to this day.
Seth gained access to anonymous searches on Google, Facebook, X (Twitter) and the most visited porn sites and for four years studied and analyzed this huge database of human secrets. The result is one of the most groundbreaking studies of the human psyche in history and how we relate to sex and relationships, politics and religion, violence and crime, racism and prejudice.
Why do we get addicted to websites and games? What happens between 2 and 4 am? How does the sentence “My husband is…” most often end? Who is more generous in sex – men or women? What can we learn from our doubles?
All data that I find far more interesting than anything AI will tell her.
Why?
Because AI doesn’t just “hallucinate” and do it too effectively and too convincingly – AI lies. And with fake news becoming a global trend as democracy declines, the risk of its lies becoming more numerous and harder to discern is not to be underestimated.
68% of users can no longer distinguish AI-generated from human content
Trust in Digital Media 2024, reuters
Yes, artificial intelligence (AI) is already playing and will continue to play a significant role in transforming marketing activities and professions in 2025. Predictions suggest that AI will replace or automate a host of routine tasks, leading to significant changes in how marketing strategies are carried out and what professions will be in demand, but I’d personally consider whether it’s worth saving money to create exactly the marketing and business strategy for your company. Or to trust an expert like Seth or someone with 20-30 years of experience who is able to sift the truth from the falsehood in the blink of an eye and what would actually work for a team or company or not.
Light in the tunnel
Ten years ago, when my son started school (2014), I chose not to be intimidated by technology, by “kids standing in front of their phones all day” and not to blame “new age” for not reading. I chose to make technology a tool to better educate and develop my child, to work, educate and support more effectively. And to share our experiences and results with others. I created DigitalKidZ, a public benefit foundation that teaches our children critical thinking, creativity, intuition, and develops their emotional intelligence.
I personally believe in technology, but I don’t trust it blindly.
I believe that AI, like all other technologies, is a tool, not a substitute for human creativity, and the challenge for us will be how to use it to enrich, not replace, the human experience.
When we learn, we will be able to educate our children to do so as well, and not be “digitally dependent and addicted”.
That is why I invest my experience and skills in the future of the business and the future of the children and I believe that the future, although sometimes scary, is not a foregone conclusion. Our future – ours, our children’s, our businesses’ – will be what we choose to make it – with the help of AI, but guided by human values.
