Collaboration across different sectors is key, as is transparent communication about AI’s potential and limitations – especially to investors and the public.
By embracing these lessons, we can create a sustainable and impactful future for AI that truly benefits society.
Let’s address the big question – are we currently headed toward an AI winter?
Are we headed for an AI winter now?
It appears that progress in AI has slowed down a bit after an explosive 2023, both with regard to new technologies released, updates to existing models and hype around generative AI.
People like Gary Marcus believe that the big leaps forward in AI model performance are becoming less frequent.
The lack of breakthroughs in generative AI and new model developments from the leaders in the space suggests a potential slowdown in progress.
Judging by investor calls, mentions of AI have also decreased, leading more to believe that the productivity gains that generative AI promised would not manifest more than what has already been achieved.
Admittedly, it isn’t much. The ROI isn’t great. Many companies struggle to find the productivity returns expected from their AI investments.
The rapid advancements and excitement around tools like ChatGPT have inflated expectations about their capabilities and potential impact.
Something previously apparent to only a small fraction of the population, mostly AI researchers, is now becoming general knowledge – large language models (LLMs).
These models face major limitations, including hallucinations and a lack of true understanding, which reduces their practical impact.
People are realizing that these technologies, when misused, are already harming the web. AI-generated content has spread across the web, from social media comments to posts, blogs, videos and podcasts.
Authentic human-generated content is becoming scarce. Future AI models will inevitably be trained on synthetic content, making it impossible to avoid and leading to worse performance over time.
We haven’t even addressed the ease of hacking generative AI, ethical issues in sourcing training data, challenges in protecting user data and many other problems that tech companies often overlook in AI discussions.
Still, some signs point against an impending AI winter in the short term.
AI technology continues to evolve rapidly, with open-source models rapidly catching up to closed models and innovative applications like AI agents emerging.
Furthermore, AI is being integrated into various industries and applications, often seamlessly (sometimes not – looking at you, AI Overviews), demonstrating at least some practical value.
It’s unclear whether these implementations will meet the tests of time.
Ongoing investment in companies like Perplexity shows investors’ confidence in AI’s potential for search, despite skeptics debunking some of the company’s claims and questioning its tactics around intellectual property.
Dig deeper: Google AI Overviews are an evolution, not a revolution
The future of AI in search and your role in it
AI is undoubtedly here to stay. My fellow automation enthusiasts and I are thrilled that everyone is now excited about this technology and exploring it themselves.
It’s important not to let the current excitement raise your expectations too high. The technology still has limits and a long way to go before reaching its full potential.
Beware of tech bros and CEOs promising uncanny ROI or sharing their doomsday predictions of the day (always so, so soon) where there will be AGI and you will be replaced by AI.
While automation is revolutionizing the workforce, change is gradual.
Progress is being made toward AGI, but reputable AI researchers believe this reality will not come in the immediate future. Numerous obstacles must still be overcome to achieve this.
Understanding any emerging technologies (especially those so widely discussed as AI is at the moment) and how they work is crucial to creating strategies that stand the test of time.
What we might see happening (in search, in particular) is one of two scenarios.
Progress continues
Implementations stand the test of time, and models improve.
For search marketers, this might mean more AI-generated content to outcompete but also improved search systems and AI-detection algorithms, easing this task by amplifying human-written, authentic voices.
Investors win. Big tech wins. Everyone wins.
That is if we solve the challenges related to ethics, security, IP and resource use. But I digress.
Progress stalls
Systems become worse. Think:
No improvement in Google AI Overviews.
Even more spam in web results.
Misinformation.
Entirely poisoned social media feeds, online forums and other digital spaces.
In this scenario, big tech will start bleeding money rapidly. (Some evidence suggests this trend has already begun.)
AI systems are, at the end of the day, expensive to develop, maintain and improve.
Failing to do so, however, will tarnish investor trust and they will eventually bow down to scaling back implementations in the area.
The public failure of several of these technologies to meet expectations will lead to the widespread loss of trust in the potential of generative AI.
In both scenarios, the brand, the authenticity of the company and its people and the approach to consumer relationships will become even more important.
The second scenario will also amplify the consumer desire for authentic non-digital experiences.
My advice to search marketers is to stay aware of the risks of AI and learn how different models work. What are their benefits and limitations? What tasks do they handle well or poorly?
Experiment with tools to boost your productivity. Many models aren’t yet ready for full marketing use, and treating them as such can worsen the issues mentioned in this article.
Dig deeper: How AI will affect the future of search
Contributing authors are invited to create content for Search Engine Land and are chosen for their expertise and contribution to the search community. Our contributors work under the oversight of the editorial staff and contributions are checked for quality and relevance to our readers. The opinions they express are their own.