Kozminski Techblog

Reading Time: 2 minutes

Artificial Intelligence (AI) is rapidly becoming an essential tool for decision-making across industries. From predictive analytics in finance to automated hiring systems, AI is increasingly influencing business strategies, government policies, and even our personal lives. While AI-driven decision-making offers speed, efficiency, and data-backed insights, it also raises an important question—are we becoming too reliant on AI to make critical choices for us?

The Rise of AI Decision-Making

AI has transformed industries by helping organizations analyze data, predict outcomes, and optimize operations at an unprecedented scale. Some of the most common applications include:

  •  Healthcare: AI-powered diagnostics assist doctors in detecting diseases faster and with higher accuracy.

  •  Finance: Banks and investment firms use AI to detect fraud and make high-frequency trading decisions.

  •  Human Resources: AI-driven recruitment software screens job candidates based on predefined criteria.

  •  Legal: AI tools predict case outcomes, analyze contracts, and assist in legal research.

In theory, AI enhances objectivity by making decisions based purely on data rather than human emotions or biases. However, the reality is far more complex.

The Illusion of Objectivity

Many believe that AI decisions are neutral and unbiased, but the truth is that AI reflects the biases of the data it’s trained on. A famous example is Amazon’s AI-driven hiring tool, which was scrapped after it was found to favor male candidates over female applicants—because it was trained on historical hiring data that was predominantly male-dominated.

Similarly, predictive policing systems in the US have been criticized for reinforcing racial biases, leading to disproportionate law enforcement actions against minority communities. When AI learns from flawed data, it doesn’t eliminate bias—it amplifies it.

When AI Gets It Wrong

While AI has made remarkable strides, it is far from perfect—and mistakes in decision-making can have serious consequences:

  •  Healthcare Errors: An AI misdiagnosing a medical condition could lead to incorrect treatment.

  •  Financial Risks: AI-driven stock trading algorithms have been responsible for flash crashes—sudden market plunges caused by automated decisions.

  •  Legal and Ethical Issues: AI used in sentencing guidelines has faced backlash for producing unfair results that disproportionately impact marginalized groups.

These failures highlight the risk of over-relying on AI without human oversight. When AI makes a mistake, who takes responsibility?

The Balance: AI as a Decision Assistant, Not a Decision Maker

The key to responsible AI adoption is not to replace human decision-makers but to empower them. Companies and policymakers need to focus on hybrid intelligence, where AI provides insights, but humans remain in control of the final decision.

How to Use AI Responsibly in Decision-Making

  1.  AI Transparency: Companies should disclose how AI makes decisions to prevent “black box” algorithms.

  2.  Human Oversight: AI should support, not replace human judgment, especially in high-risk sectors like healthcare and law.

  3.  Bias Audits: Organizations must regularly audit AI systems to detect and correct biases.

  4.  Ethical AI Guidelines: Governments and corporations should work together to create ethical AI regulations that ensure fairness and accountability.

Final Thoughts: Who Should Have the Final Say?

AI decision-making is a powerful tool, but it should never replace human intuition, ethics, and accountability. While AI can process vast amounts of data, it lacks the ability to understand context, morality, and human emotions—factors that often influence real-world decisions.

Sources:

The AI Ethics Guidelines from the European Commission

Harvard Business Review: The Risks of AI Decision-Making

MIT Sloan Review: AI Bias and Decision-Making

The World Economic Forum: Balancing AI and Human Judgment

Stanford AI Index Report (2023)

written with help of Grok

Reading Time: 2 minutes

Artificial Intelligence (AI) is revolutionizing how businesses operate, streamlining workflows, and improving efficiency across industries. From automated customer service chatbots to AI-powered financial risk analysis, we are witnessing a shift towards an increasingly intelligent and automated workforce. However, while the benefits of AI-driven automation are undeniable, there are significant challenges that often go unnoticed. Are we truly prepared for the consequences of full-scale AI adoption?

AI and the Illusion of Efficiency

At first glance, AI appears to be a game-changer for productivity. Algorithms can process vast amounts of data in seconds, eliminating human error and improving decision-making. Yet, studies indicate that automation does not always lead to seamless efficiency. Many companies find themselves facing implementation fatigue—a phenomenon where employees struggle to adapt to rapid technological changes, often leading to lower productivity rather than higher output.

Additionally, AI’s reliance on predictive models can sometimes reinforce existing inefficiencies instead of resolving them. For example, HR departments using AI to filter job candidates may inadvertently reinforce biases in recruitment, selecting candidates based on past hiring patterns rather than merit.

The Psychological Toll of AI on Workers

AI doesn’t just change the way we work—it changes how we feel about work. Recent surveys suggest that over 60% of employees experience heightened anxiety due to AI-driven changes in their workplaces. The uncertainty of job displacement or the constant need to upskill to stay relevant can create a stressful work environment.

Interestingly, young professionals under 30 seem to be the most affected, with 72% reporting concerns about being replaced by AI within the next decade. While automation promises to eliminate repetitive tasks, it also removes a sense of job security, leaving many workers questioning their long-term career prospects.

AI Governance and the Ethics Dilemma

The rapid expansion of AI raises important questions about ethics and governance. Who is responsible when an AI system makes a mistake? If an autonomous vehicle causes an accident or a healthcare AI misdiagnoses a patient, where does the liability fall?

Governments and organizations are scrambling to introduce regulatory frameworks that ensure AI accountability, but the pace of regulation is lagging behind technological advancements. Without transparent guidelines, AI adoption could lead to serious ethical and legal complications that businesses are ill-equipped to handle.

So, What’s Next?

AI-driven automation is here to stay, but companies must rethink their strategies to ensure that these systems enhance human work rather than replace it entirely. Instead of a race toward full automation, businesses should focus on hybrid intelligence—a model where AI supports human decision-making rather than replacing it.

Moreover, organizations should prioritize AI literacy to prepare employees for the future workplace. Investing in reskilling programs and fostering adaptability will be key to ensuring that AI is an asset, not a source of anxiety.

Conclusion

The AI revolution presents both exciting opportunities and significant challenges. While automation holds the potential to reshape industries for the better, it also raises critical concerns about efficiency, job security, and ethical accountability. The key to sustainable AI integration lies in balanced implementation, ensuring that technology serves humanity rather than the other way around.

Are we ready for the next wave of AI automation? The answer depends on how we choose to shape its impact today.

Sources:

The World Economic Forum’s Future of Jobs Report (2023)

MIT Technology Review: How AI Is Changing Work

Harvard Business Review: The Productivity Paradox of AI

McKinsey’s AI and Automation Report

Stanford University’s AI Index Report (2023)

written with help of google gemini

Reading Time: 2 minutes

Introduction to Prompt Engineering in Hiring

Prompt engineering is a technique in artificial intelligence (AI) in which particular instructions or “prompts” are designed to produce exact outputs from generative models. This method has gained popularity across sectors for activities such as automating content production, optimizing procedures, and hiring.

In recruiting, quick engineering may help firms evaluate resumes, create job descriptions, and undertake early applicant assessments. Businesses may use targeted prompts to enhance the accuracy and effectiveness of their recruiting procedures, saving time and resources.

InPost’s approach represents one of the earliest large-scale adoptions of prompt engineering in recruitment within the logistics sector.

InPost’s Practical Application of Prompt Engineering

InPost, led by Rafał Brzoska, has implemented fast engineering to improve recruiting for logistics and technology professions. They’ve used powerful generative AI technologies like Zapier and proprietary GPT-like solutions to automate screening and evaluating candidates’ core abilities by comparing applications to job-specific requirements.

Recruitment Optimization:

InPost optimizes recruitment by using AI to filter resumes and better match individuals to jobs. Using language models like as OpenAI’s GPT, the organization compares applicants’ talents, experiences, and qualifications to established job criteria. The AI creates score measures that allow recruiters to focus on top applicants without the need for human filtering. For example, prompts are intended to collect important facts, check alignment with job tasks, and even indicate potential mismatches, therefore expediting the hiring process.

Enhanced Candidate Communication:

Another innovative application is in candidate interaction. Using tools like GPT-based AI or conversational platforms such as Intercom or Drift (integrated with generative AI), InPost ensures that applicants receive personalized updates at each recruitment stage. This includes AI-generated emails with tailored feedback, which improves transparency and builds trust with candidates.

Training and Development of AI Models:

InPost likely employs fine-tuning methods on models like GPT, using historical hiring data to enhance the accuracy of candidate evaluations. These customizations ensure the AI recognizes industry-specific terminology and evaluates nuanced qualifications critical in logistics roles. By integrating platforms like Make.com or Zapier, InPost further connects its generative AI tools to broader workflow automations, creating a seamless recruitment system.

Insights from Rafał Brzoska

Rafał Brzoska has highlighted how leveraging AI, including prompt engineering, has significantly reduced time-to-hire and improved the overall quality of recruits at InPost – On his LinkedIn, he emphasizes that embracing such technologies is not just a trend but a necessity in staying competitive in the logistics and tech-driven market. The company’s AI initiatives, including those in hiring, reflect their commitment to efficiency and excellence in operations

Broader Implications for Other Companies
Other organizations can take inspiration from InPost’s model by adopting prompt engineering for:

1. Crafting detailed, unbiased job descriptions that attract diverse talent.
2. Pre-screening resumes using AI to filter out mismatches.
3. Using AI-driven chatbots to conduct initial candidate interviews.

________________________

Written with help of Claude.ai

Sources:

https://www.pracuj.pl/praca/prompt-engineer-warszawa-czerniakowska-87a,oferta,1002513338

https://www.ey.com/en_lb/weoy/class-of-2022/poland

https://inpost.pl/kariera/data-ai

https://www.linkedin.com/jobs/view/ai-engineer-at-inpost-4083473735/

InPost oferuje pracę przyszłości? Prompt engineer będzie pracował m.in. z GPT-4

Reading Time: 2 minutes

With breakthroughs in AI, organization now have a unique chance to alter the way they run consumer testing by leveraging sentiment analysis to acquire a richer, more accurate understanding of customer input. This method expands on standard surveys and focus groups by allowing AI to decipher emotional reactions and even assess the legitimacy of customer replies.

Why It’s a Game-Changer for Companies ?

AI-powered sentiment analysis, combined with speech-to-text capabilities, provides a valuable instrument. It catches every detail in a consumer’s voice, from eagerness and hesitancy to confidence and uncertainty, giving businesses insight into the underlying emotions driving consumer behaviors. AI can transform verbal input into useful insights by identifying alterations in tone or word choice that convey deeper sentiment.

One of the most amazing features of AI in consumer tests is its prediction ability to determine the validity of replies. AI can distinguish between real and filtered or exaggerated feedback using data patterns and algorithms that identify consistency, spontaneity, and natural expression. This discovery enables businesses to focus on true insights while eliminating biases that might distort outcomes.

Mariot International Case Study:

Marriott International uses AI-powered sentiment analysis to handle customer reviews and comments from over 7,000 hotels worldwide. By evaluating themes and feelings in guest evaluations, including as room cleanliness, staff friendliness, and amenity quality, Marriott may identify opportunities for improvement at individual properties. For example, if a hotel receives recurrent complaints about service delays, managers may rapidly address the problem by changing staffing or processes.

EA Case Study:

Leading video game publisher Electronic Arts (EA) uses artificial intelligence (AI) sentiment analysis to handle player reviews and comments for their titles. EA determines which aspects users like and which ones could require a little more work by examining sentiment at the game level.

This kind of potentially ignored information can help with problem patches, game upgrades, and the creation of new games. EA can modify a game’s micro-transaction system to make it more player-friendly and enhance reviews from third-party critics if feedback analysis shows that users are dissatisfied with the system, which is a very common problem in the video game industry. It basically means that they analyze your in game chats.  

AI sentiment analysis is also relevant in a variety of industries, including automotive, healthcare, e-commerce, and entertainment. Businesses that regularly monitor and respond to consumers maintain a great online reputation, develop customer loyalty, and gain a competitive advantage in their particular sectors.

Sentiment market in Poland

As more people use AI-driven products, the sentiment analysis market in Poland is expanding. Growing social media usage and the need for real-time analytics in sectors like finance, retail, and telecommunications are driving this expansion. Globally, the market is growing at a compound annual growth rate (CAGR) of more than 17%, with developments in machine learning and natural language processing (NLP) playing important roles. Applications in Poland are mostly found in brand reputation tracking and customer experience management, where businesses use local language natural language processing (NLP) techniques to increase efficiency.

_______________

Resources:

https://www.widewail.com/blog/10-real-world-examples-of-ai-topic-sentiment-analysis

sentire2024karlinska.pdf

Sentiment Analysis Software Market Report 2024 – Sentiment Analysis Software Market Trends And Overview

15 Sentiment Analysis Statistics in 2024 – Marketing Scoop

Consumer Sentiment in Poland Remains Positive

_______________

Text is basing on the output of Chat GPT

Reading Time: < 1 minute

As AI-powered technologies advance, companies can automate a variety of operations that formerly required a specialized sales crew with relatively little knowledge. This article will briefly present a strategic approach of building AI powered sales system with tools that are available for everyone.

AUTOMATED LEAD GENERATION WITH APOLLO.IO

This tool enriches each prospect’s profile with verified contact details, including email addresses, phone numbers and social media links allowing for more effective and informed outreach. Integration of Apollo.ai with platforms like Airtable (covered next) creates a seamless workflow for managing and organizing leads within the pipeline.

ORGANIZING LEADS IN AIRTABLE

Airtable serves as flexible CRM-like (customer relationship managament system) database that is ideal for tracking leads, managing contact information and visualising current marketing efforts. Its automation features simplify many repeating task like reporting, status updates etc.

OUTREACH AUTOMATION WITH INSTANTLY.AI

This tool is a core of whole direct outreach. By automating follow-up sequences based on recipient engagement, Instantly.ai ensures continous outreach without manual effort. Tool offers tailored messaging and timely follow-ups within fully automated marketing campaigns. It serves as a scalable solution that mimics personalized interactions across numerous leads efficiently.

SYSTEM INTEGRATION WITH ZAPIER.COM

Zapier automates workflows by seamlessly connecting various platforms, enabling synchonized data updates and task triggers across tools like those mentioned above. For example. Zapier can initiate email sequences or update lead statuses based on present actions, streamlining operations without manual intervention.

Integrating mentioned tools unlocks many opportunities for scalable growth and seamless automation for smaller to medium businesses, that with this knowledge won’t even need a marketing agency.

P.S This automation serves as a basis for my company to generate 5-6K zł profit every month for 7 consecutive months!

__________________

Reference links:

Apollo.io

Airtable.com

Instantly.ai

Zapier.com

__________________