Baylis & Driver (Nature Neuroscience, 2001) have recently presented data on the response of neurons in macaque inferotemporal cortex (IT) to various stimulus transformations. They report that neurons can generalize over contrast and mirror reversal, but not over figure-ground reversal. This finding is taken to demonstrate that ``the selectivity of IT neurons is not determined simply by the distinctive contours in a display, contrary to simple edge-based models of shape recognition'', citing our recently presented model of object recognition in cortex (Riesenhuber & Poggio, Nature Neuroscience, 1999). In this memo, I show that the main effects of the experiment can be obtained by performing the appropriate simulations in our simple feedforward model. This suggests for IT cell tuning that the possible contributions of explicit edge assignment processes postulated in (Baylis & Driver, 2001) might be smaller than expected.

Across two essays, I explore how artificial intelligence (AI) applications can help businesses automate customer service with deep learning-driven natural conversation and improve employee performance with work supervision. I apply machine learning methods such as audio analytics and text mining, as well as field experiments to explore these new AI-driven capabilities in customer service and employee supervision automation. Substantively, this research tackles emerging business questions regarding how AI applications can assist customer purchases and employee job performance. In Essay One, I apply two experiments to investigate when and how AI voicebots work or struggle in persuading customers relative to human agents. In Experiment 1, I apply audio analytics to extract agents’ voice features (i.e., pitch, amplitude, and speed) and speech content (i.e., selling adaptivity). My analyses suggest two distinct routes to explain how agents’ speech patterns account for their performance. Analyses in Experiment 2 demonstrate that relative to human agents, AI bots could backfire and lead to worse performance when the customer persuasion task is more complex. In my second essay, I explore the coexistence of performance improvement and employee resistance to AI supervision. Specifically, I develop a novel two-by-two field experiment, which randomly assigns the AI or human supervision entity and discloses the entity or not, to separate the economic gain from negative reactance to AI. In addition, I uncover the underlying mechanism by identifying employees’ subjective bias to the AI feedback quality and heightened fear of job replacement once they know the supervision entity is AI rather than human managers. I propose two strategies to alleviate employees’ resistance to AI supervision. / Business Administration/Marketing

Denna studie syftar till att undersöka hur människor som skapar och implementerar AI-algoritmer förhåller sig till etiska dilemman som de möter i samband med sitt arbete. Målet var att fastställa vilka etiska dilemman som finns, svårigheter och åsikter kopplade till dessa samt hur de kan hanteras. Studien bygger på 5 semistrukturerade kvalitativa intervjuer med IT-arkitekter som berörs av AI i sitt dagliga arbete, samt en litteraturgenomgång. Litteraturgenomgången tydliggör begreppen: AI, AI-algoritmer och etik, redogör för fastställda etiska principer kopplade till AI och beskriver människans relation till AI och etik. Intervjuerna ämnar att upplysa om hur arbetet ser ut i praktiken och förklara hur etiska dilemman kan förekomma vid skapandet och implementeringen av AI-algoritmer, resultatet analyseras utifrån litteraturgenomgången. Resultatet visar att människan har en viktig, men samtidigt svår roll när det gäller att ta hänsyn till etik i samband med sitt arbete. De etiska principerna är alla av stor vikt och intresse, de skiljer sig dock från varandra i förhållande till komplexitet och relevans. Människan bakom maskinen är medveten om svårigheterna med sitt dagliga arbete och försöker efter bästa förmåga tillämpa samtliga etiska principer i det. Det är idag inte möjligt att fullkomligt följa de etiska principerna. Detta skapar ett behov av beslutsfattande gällande till vilken grad teknologins effektivitet kan utnyttjas utan att skapa allt för stora etiska risker i samhället.

Die vorliegende Arbeit untersucht die Verwendung von Large Language Models (LLMs) in Penetrationstests von Web-Anwendungen. Ziel ist es, die Arbeit von Penetrationstestern zu unterstützen und den Prozess zu beschleunigen, um Sicherheitslücken in Web-Anwendungen effektiver aufzudecken und zu beheben. Die Arbeit vergleicht verschiedene Ansätze und prüft, wie LLMs wie ChatGPT und andere die Effizienz des Penetrationstests verbessern können. Es wird evaluiert, ob durch die Anwendung von LLMs der notwendige Aufwand für Penetrationstests reduziert werden kann, um Sicherheitslücken in Web-Anwendungen effektiver aufzudecken und zu beheben. Die Arbeit leistet einen Beitrag zum Thema, indem sie die Möglichkeiten und Grenzen von LLMs im Kontext der Penetrationstestung untersucht, bewertet und den aktuellen Stand skizziert.:1 Intro 2 Basics 2 1 Web Application Security 2 2 Penetration Testing 2 3 Penetration Testing Standards 2 4 Penetration Testing Tools 2 5 Artificial Intelligence 2 6 Large Language Models 2 7 LLM prompting techniques 2 8 AI’s Growing Role in Cybersecurity 2 9 Penetration Testing and AI 2 10 Research Objectives and Scope 2 11 Significance of the Study and Research Question 2 12 Structure of the Thesis 3 Literature Review 4 Market Analysis 4 1 Use of LLMs in Combination with Existing Penetration Testing Software 4 2 Open-Source Solutions Leveraging LLMs 4 3 Commercial Solutions Leveraging LLMs for Cybersecurity purposes 4 4 ChatGPT-GPTs 4 5 Identifying the Need for Optimization in Penetration Testing Processes 4 6 Opinions of Penetration Testers on Generative AI Use 5 Methodology 5 1 Research Methods and Approaches 5 2 Benchmarks Used for Evaluation 6 Concept and Implementation 6 1 Limitations of LLMs 6 2 Deciding Which LLM Models to Use 6 3 Identifying and Executing Tasks with LLMs 6 4 Tailoring the LLM for Penetration Testing 6 5 Resource Requirements 7 Evaluation of LLMs for Penetration Testing 7 1 Interviews: Identifying the use of LLMs for Pentesting 7 2 Preparing the Test Environment 7 3 Evaluation of Command Generation 7 4 ChatGPT Assistant GPT 7 5 Google Gemini Advanced 7 6 Discussion of results 7 7 Answering the Research Question 7 8 Resulting Penetration Testing Workflow 8 Conclusion / The thesis examines the use of Large Language Models (LLMs) in web application penetration testing. The goal is to support penetration testers and accelerate the process, to identify and fix security vulnerabilities in web applications more effectively. The thesis compares different approaches and evaluates how LLMs, such as ChatGPT and others, can improve the efficiency of penetration testing. It is evaluated whether the application of LLMs can reduce the necessary effort for penetration testing, to more effectively identify and fix security vulnerabilities in web applications. The research contributes to the topic by investigating, evaluating, and outlining the possibilities and limitations of LLMs in the context of penetration testing.:1 Intro 2 Basics 2 1 Web Application Security 2 2 Penetration Testing 2 3 Penetration Testing Standards 2 4 Penetration Testing Tools 2 5 Artificial Intelligence 2 6 Large Language Models 2 7 LLM prompting techniques 2 8 AI’s Growing Role in Cybersecurity 2 9 Penetration Testing and AI 2 10 Research Objectives and Scope 2 11 Significance of the Study and Research Question 2 12 Structure of the Thesis 3 Literature Review 4 Market Analysis 4 1 Use of LLMs in Combination with Existing Penetration Testing Software 4 2 Open-Source Solutions Leveraging LLMs 4 3 Commercial Solutions Leveraging LLMs for Cybersecurity purposes 4 4 ChatGPT-GPTs 4 5 Identifying the Need for Optimization in Penetration Testing Processes 4 6 Opinions of Penetration Testers on Generative AI Use 5 Methodology 5 1 Research Methods and Approaches 5 2 Benchmarks Used for Evaluation 6 Concept and Implementation 6 1 Limitations of LLMs 6 2 Deciding Which LLM Models to Use 6 3 Identifying and Executing Tasks with LLMs 6 4 Tailoring the LLM for Penetration Testing 6 5 Resource Requirements 7 Evaluation of LLMs for Penetration Testing 7 1 Interviews: Identifying the use of LLMs for Pentesting 7 2 Preparing the Test Environment 7 3 Evaluation of Command Generation 7 4 ChatGPT Assistant GPT 7 5 Google Gemini Advanced 7 6 Discussion of results 7 7 Answering the Research Question 7 8 Resulting Penetration Testing Workflow 8 Conclusion

No description available.

No description available.

No description available.

No description available.

The Copyright law is the legal part of the intellectual property that is considered to be updated and adjustable enough to handle a technological development. As artificial intelligence develops to be more autonomous and able to create works independently, the copyright law becomes subject to new requirements and challenges. The essay’s purpose is to investigate the legal situation regarding copyright and artificial intelligence from both a national and an international point of view in order to describe what possibilities works made by artificial intelligence have to get copyright protection. The essay has been initially based on Swedish copyright law to clarify what the swedish law requires a common work but particularly a computer generated work in order for the work to get copyright protection. Further, the essay has had a comparative method in the section regarding international copyright. A study has been made from a EU legal perspective to compare copyright law both on a EU level but also a national level, and to investigate the harmonizing effect of various directives that have been relevant to the essay’s issues. In order to give an as fair as possible answer to the questions the essay seeks to answer, an investigation has been made also of the british and american copyright law. In connection with this, a more comprehensive mapping of international copyright has also been relevant in order to obtain a basis for proposals on what mainly the national copyright law could or should be (de lege ferenda). In the light of swedish copyright law, artificial intelligence and computer programs have also been taken into consideration in order to investigate whether AI systems can be equated with computer programs and thus get copyright protection. Before the essay’s final reflections an investigation has also been made regarding copyright infringement in order to judge who’s responsible or what person that should be responsible in the case of a copyright infringement made by artificial intelligence.

This research extends a tradition of distributed theories of mind into the implementation of a distributed problem solver. In this problem solver a number of ideas from Minsky's Society of Mind are implemented and are found to provide powerful abstractions for the programming of distributed systems. These abstractions are the cauldron, a mechanism for instantiating reasoning contexts, the frame, a way of modularly describing those contexts and the goal-node, a mechanism for bringing a particular context to bear on a specific task. The implementation of both these abstractions and the distributed problem solver in which they run is described, accompanied by examples of their application to various domains.